Arabinose content of extracellular polysaccharide plays a role in cell aggregation of Azospirillum brasilense


  • Efrat Bahat-Samet,

    1. Department of Plant Pathology and Microbiology and The Otto Warburg Minerva Center for Agricultural Biotechnology, The Hebrew University of Jerusalem, Faculty of Agricultural, Food and Environmental Quality Sciences, P.O. Box 12, 76100 Rehovot, Israel
    Search for more papers by this author
  • Susana Castro-Sowinski,

    1. Department of Plant Pathology and Microbiology and The Otto Warburg Minerva Center for Agricultural Biotechnology, The Hebrew University of Jerusalem, Faculty of Agricultural, Food and Environmental Quality Sciences, P.O. Box 12, 76100 Rehovot, Israel
    2. Departamento de Bioquímica, Instituto Clemente Estable, Unidad Asociada de Bioquímica, Facultad de Ciencias, Universidad de la República, Av. Italia 3318, 11600 Montevideo, Uruguay
    Search for more papers by this author
  • Yaacov Okon

    Corresponding author
    1. Department of Plant Pathology and Microbiology and The Otto Warburg Minerva Center for Agricultural Biotechnology, The Hebrew University of Jerusalem, Faculty of Agricultural, Food and Environmental Quality Sciences, P.O. Box 12, 76100 Rehovot, Israel
      *Corresponding author. Tel.: +972-8-948-92-16; fax: +972-8-946-67-94, E-mail address:
    Search for more papers by this author

*Corresponding author. Tel.: +972-8-948-92-16; fax: +972-8-946-67-94, E-mail address:


Extracellular polysaccharides play an important role in aggregation and surface colonization of plant-associated bacteria. In this work, we report the time course production and monomer composition of the exopolysaccharide (EPS) produced by wild type strain and several mutants of the plant growth promoting rhizobacterium (PGPR) Azospirillum brasilense. In a fructose synthetic medium, wild type strain Sp7 produced a glucose-rich EPS during exponential phase growth and an arabinose-rich EPS during stationary and death phase growth. d-glucose or l-arabinose did not support cell growth as sole carbon sources. However, glucose and arabinose-rich EPSs, when used as carbon source, supported bacterial growth. Cell aggregation of Sp7 correlated with the synthesis of arabinose-rich EPS. exoB (UDP-glucose 4′-epimerase), exoC (phosphomannomutase) and phbC (poly-β-hydroxyburyrate synthase) mutant strains, under tested conditions, produced arabinose-rich EPS and exhibited highly cell aggregation capability. A mutant defective in LPS production (dTDP 4-rhamnose reductase; rmlD) produced glucose-rich EPS and did not aggregate. These results support that arabinose content of EPS plays an important role in cell aggregation. Cell aggregation appears to be a time course phenomenon that takes place during reduced metabolic cell activity. Thus, aggregation could constitute a protected model of growth that allows survival in a hostile environment. The occurrence of exoC and rmlD was detected in several species of Azospirillum.


Bacteria of the genus Azospirillum are free living nitrogen fixing rhizobacteria that are found in close association with plant roots. Azospirilla are able to promote plant growth through improvement of root development. Evidence suggests that production of phytohormones by the bacteria play an important role in plant growth promotion (PGP) [1]. Beneficial effects result in increased crop yield in many plants of agronomic importance.

The Azospirillum–plant interaction is not yet fully understood, but it is known that bacterial surface components, such as extracellular polysaccharides and proteins, are involved in plant colonization [2–4]. Among different Azospirillum species, A. brasilense is the best studied. A. brasilense cells have the ability to aggregate and flocculate, and these properties may positively affect survival in soil [5]. The phenomenon of bacterial aggregation is of great interest in production, storage and survival of bacterial inoculants for agriculture application. The concentration of exopolysaccharides (EPS) produced by different strains of A. brasilense strongly correlates with their extent of aggregation [6]. In addition, Burdman et al. [7] suggested that the arabinose present in the EPS and in the capsular polysaccharide (CPS) produced by A. brasilense play an important role in determining its aggregation capability.

In this work we have studied the time course production and composition of EPS produced by A. brasilense wild type strain Sp7 and several mutant strains: exoB (UDP-glucose 4′-epimerase) and exoC (phosphomannomutase) mutants, that are known to fail in production of high-molecular-weight EPS and show an altered distribution of the EPS component [8,9]; a phbC mutant (poly-β-hydroxyburyrate synthase), which does not accumulate poly-β-hydroxyburyrate (PHB), overproduce EPS and shows increased cell aggregation capability [10]; and a rmlD mutant (dTDP 4-rhamnose reductase) with a modified lipopolysaccharide (LPS) production that shows a pleiotropic phenotype with altered colony morphology, increased EPS production, impaired attachment to maize roots and diminished maize root colonization [11]. In this work, we report that EPS properties and composition varies during time course growth of the mutant strains as compared to the wild type. In addition, we support previous findings by Burdman et al. [7] in that EPS arabinose content strongly correlates with cell aggregation capability. It is also shown that A. brasilense is able to metabolize EPS of its own production as a sole carbon source. Finally, we studied the occurrence of exoB, exoC, and rmlD genes in several species of Azospirillum.

2Materials and methods

2.1Bacterial strains and growth conditions

The A. brasilense strains used in this study are shown in Table 1. For aggregation assays and EPS extraction, bacteria were grown in 250 ml Erlenmeyer flasks, with 50 ml liquid medium containing fructose as carbon source and high C:N ratio as previously described [6]. Flasks were inoculated with exponential phase cultures at an initial OD540 of approximately 0.05 (about 107 CFU ml−1), and incubated on a rotary shaker (150 rpm) at 30 °C. Kanamycin was added at 50 μg ml−1 for mutant strains.

Table 1.  Bacterial strains used in this study
StrainRelevant characteristicReference or origin
A. brasilense Sp7 (ATCC 29145) (closely related to Cd)Wild-type strainTarrand et al. [12]
A. brasilense exoB mutantSp7 exoB::Tn5 mutantMichiels et al. [9]
A. brasilense exoC mutantSp7 exoC::Tn5 mutantMichiels et al. [9]
A. brasilense phbC mutantSp7 phbC::Km mutantKadouri et al. [10]
A. brasilense rmlD (LPS mutant)Cd rmlD::Tn5 mutant (with modified LPS)Jofré et al. [11]
A. lipoferum (ATCC 29708)Wild-type strainTarrand et al. [12]
A. amazonense (ATCC 35119)Wild-type strainMagalhaes et al. [13]
A. halopraeferens (ATCC 43709)Wild-type strainReinhold et al. [14]
A. irakense KBC1Wild-type strainKhammas et al. [15]
A. doebereinerae GSF-71TWild-type strainEckert et al. [16]

2.2Extraction of EPS, analysis of sugar content by high performance liquid chromatography (HPLC) and aggregation assays

Strains were grown for 24, 48, 72 and 96 h as described above. Polysaccharides were extracted by fractionation with cold ethanol as described by Del Gallo and Haegi [17] with some modifications [7]. Evaluation of the sugar amount was done by the anthrone method, using glucose as a standard [18]. Microbial mass was determined by measurement of dry cell mass at 80 °C until a constant weight.

Following polysaccharide extraction, samples were dialyzed against distilled water (during three days at 4 °C, the cut-off the membrane was 12,000–14,000), lyophilized and then hydrolyzed in 0.5 M H2SO4 at 90 °C for 18 h. After neutralization with Ba(OH)2 and centrifugation (7000g, 15 min, 4 °C), the supernatant was filtered through 0.45 μm membranes and 20 μl samples were loaded onto a Kentron HPLC apparatus coupled with constant pulse detection using a sugar pak™ 1 column. Double distilled water and EDTA (50 mg l−1) were used as eluants at a flow rate of 0.5 ml min−1. d-glucose and l-arabinose, which are two of the main components of the extracellular polysaccharides of A. brasilense grown on fructose [4], were well resolved using this procedure. l-rhamnose and d-galactose, which are also two from the main components of the EPS, could not be separated by this method and the results are shown as rhamnose plus galactose. Specific EPS production was defined as equivalent of glucose by cell dry weight (μg mg−1).

2.3Scanning electron microscopy (SEM)

Bacteria were washed by centrifugation (4000g, 20 min, twice), resuspended in 10 mM phosphate buffer (pH 7.2), fixed in 2% (v/v) glutaraldehyde for 3 h, dehydrated through a graded ethanol series and critical point drier. Dried bacteria and EPS were coated with gold by a diode sputtering system (Polaroid E5000) and observed with a JEOL JSM-35C scanning electron microscope.

2.4Effect of purified EPS on plant growth

Seeds of wheat (Triticum aestivum cv. Atir) were surface sterilized as described by Dobbelaere et al. [19]. They were then placed on a disc of filter paper with 1.5 ml of sterile tap water in a Petri dish. The seeds were incubated for 24 h in the dark at 25 °C and then inoculated with 1 ml of washed cells (4 × 106 and 4 × 107 CFU) of overnight cultures of wild type Sp7 and 200 μg purified EPS. Gum Xanthan was used as a negative control. The number and length of roots and the length of shoots were measured four and eight days after inoculation.

2.5Occurrence of exoB, exoC and rmlD in different species of Azospirillum

Occurrence of exoB, exoC and rmlD genes on A. brasilense, A. irakense, A. doebereinerae, A. lipoferum, A. halopraeferens and A. amazonense was analyzed by PCR. Genomic DNA was purified using Wizard® Genomic Purification Kit (Promega) following manufacturer conditions. Primers were designed via the Primer3 Program ( using exoB, exoC and rmlD sequences of A. brasilense Sp7 (GenBank, Accession Nos.: Z25478, U20583 and AF349575, respectively). PCR was performed on an automated PCR thermoblock (Mastercycler gradient; Eppendorf). The PCR mixture of 25 μl contained 50 mM Tris (pH 8.3), 500 μg ml−1 bovine serum albumin (BSA), 3 mM MgCl2, 200 μM deoxynucleoside triphosphates, 1 U RedTaq™ DNA polymerase (Sigma Co.), 17 pmol of each primer, and 5 μl of DNA. Cycling conditions were as follows: 95 °C for 5 min, followed by 30 cycles at 95 °C for 30 s, 45 °C for 45 s, and 72 °C for 1 min, and a final dwell of 72 °C for 2 min. Amplified products were separated in 1% agarose gels containing 0.5 μg ml−1 of ethidium bromide and photographed using the AlphaImager 1220 v5.5 system (Alpha Innotech Corporation). PCR fragments were purified from agarose gel (QIAquick PCR purification kit; Qiagen, USA) and DNA sequencing was performed using the ABI Prism 377 DNA sequencer (Applied Biosystem Inc., CA, USA).


3.1Time course production of cell dry weight and specific EPS

The time course production of dry cell mass (biomass) and specific EPS in wild type strain Sp7 and the four mutants was followed (Fig. 1). Biomass production was slower in all mutants than in Sp7 strain. After 48 h of growth Sp7 reached a biomass 40% higher than the mutants (Fig. 1(a)). In all cases, the higher specific EPS production was detected 24 h after growth (exponential growth phase), but its content clearly decreased after that (Fig. 1(a)). The increase in Sp7 biomass production accompanied by decrease in specific EPS suggests that bacteria might use the EPS for growth and/or cell maintenance.

Figure 1.

Time course production of dry cell mass (a) and specific EPS (b) in wild type and mutant strain of A. brasilense. Results are the averages of three independent experiments. Symbols are as follows: A. brasilense Sp7 (▪), exoB mutant (•), exoC mutant (□), LPS mutant (▴) and phbC mutant (^).

3.2Time course of sugar composition of EPS

The time course of sugar composition of Sp7 and the four mutants EPSs showed that, during exponential growth phase (48 h after growth), glucose was the main sugar in Sp7 EPS. However, during stationary and death growth phases, arabinose became the main one (Table 2). Among tested sugars, arabinose was the main EPS component produced by exoB, exoC and phbC mutants, showing a slight tendency to increase during growth, probably at operating cost of glucose or because an unbalance in glucose/arabinose ratio (Table 2). Rhamnose/galactose content, in the wild type, exoB, exoC and phbC mutants, slightly (10–13%) increased only after 96 h growth. The LPS mutant showed a steady EPS composition (70% glucose, 30% arabinose) and rhamnose/galactose was not detected, as it was expected for a mutant in rhamnose biosynthesis [11].

Table 2.  Relative concentration (%) of the main monosaccharides in EPS and aggregation (%) of Sp7 and the four mutant strains growing in high C:N medium
StrainGrowth time (h)Glucose (%)Rhamnose and galactose (%)Arabinose (%)% Aggregation
  1. Results are the averages (three replicates per treatment each experiment) from three independent experiments. Different letters represent significant differences (P= 0.05) among the various sugars in all strains, as determined by one-way analysis of variance (ANOVA). Values of % aggregation are mean of three experiments plus standard error.

Sp72495a5b0e4 ± 0
 4898a1b1e23 ± 3
 7212c1b87a84 ± 2
 967d13a80a70 ± 1
Sp7 exoB::Tn52438c1b61b44 ± 2
 4824c1b75a73 ± 3
 7212c0c88a81 ± 6
 9620c8a72a75 ± 9
Sp7 exoC::Tn52476b0c24d27 ± 2
 4847c0c53c69 ± 3
 7235b11a54b87 ± 3
 9641b12a47c70 ± 9
Sp7 phbC::Km2424d1c75a66 ± 2
 485e3b92a64 ± 7
 727c2b91a73 ± 2
 966d11a83a72 ± 3
Cd rmlD::Tn52464b0c36c4 ± 1
 4868b0c32d6 ± 1
 7270a0c30d11 ± 2
 9670a0c30d14 ± 2

3.3Growth of wild type Sp7 strain with EPS as carbon source

The hypothesis that Sp7 strain support cell growth using its own EPS as carbon source was tested. Sp7 strain was grown in liquid medium supplemented with 1% (w/v) purified and dialyzed EPS from its own and from the four mutant strains production (obtained from 24 h cultures) as a sole carbon source. As it was shown above, after 24 h A. brasilense Sp7 and the LPS mutant produced a glucose-rich EPS and exoB, exoC and rmlD mutant strains produced an arabinose-rich EPS. Absence of proteins on purified EPS fractions were confirmed by Bradford method, following manufacturer conditions (BioRad). Fructose was used as positive control; Gum Arabic and Gum Xanthan were used as negative control. After 48 h growth, Sp7 reached similar biomass (approximately 0.75 absorbance units at 540 nm) when growing on fructose and on the different bacterial purified EPS. In addition, l-arabinose, d-galactose or d-glucose monosaccharides did not support bacterial growth as a sole carbon source (data not shown).

3.4Time course of cell aggregation and sugar EPS composition

A correlation between relative amounts of arabinose in EPS and cell aggregation was observed (Table 2). Apparently, 30% arabinose was needed for visible cell aggregation. In the tested conditions Sp7, showed visible cell aggregation after 72 h growth, whereas exoB, exoC and phbC mutants aggregated sooner, at 24 h of growth. The LPS mutant always showed low relative arabinose EPS concentration and non-aggregation capability. Sp7 aggregates appeared small and spherical. Meanwhile, exoB, exoC and phbC mutants produced large and like-star aggregates (not shown). In addition, these mutants stuck to the glass surface and were not easily removed by shaking as compared to Sp7 and LPS mutant cells.

3.5Scanning electron microscopy of cells and EPS

As phbC mutant produced the higher amount of stickier rich-arabinose EPS, it was chosen as a model to compare bacterial morphology and EPS appearance between both rich-glucose EPS (Sp7) and rich-arabinose EPS producer (phbC mutant) strains. SEM showed that the Sp7 cells stayed individually while phbC mutant cells appeared sticky (Fig. 2). In addition, also EPS appearance was different. Under SEM, EPS produced by Sp7 showed shorter needle-like filaments whereas EPS from the phbC mutant gave the impression of being longer needle-like filaments (Fig. 2).

Figure 2.

Scanning electron microscopy of cells and EPS from wild type strain and phbC mutant, after 24 h growth. (a) Cells of Sp7; (b) cells of phbC mutant; (c) EPS produced by Sp7 and; (d) EPS produced by phbC mutant.

3.6Phytostimulatory effect of EPS on wheat

The possibility that EPS from A. brasilense has a phytostimulatory effect or can elicit a plant response was tested in wheat seedlings. Addition of different amounts of purified EPSs from Sp7 and mutants did not show any promotion on root and shoot development, whereas the inoculation with washed cells of A. brasilense (106–107 CFU per seed) significantly increased plant growth (data not shown).

3.7Occurrence of exoB, exoC and rmlD genes in other species of Azospirillum

The occurrence of exoC and rmlD genes in other species of the genus Azospirillum was analyzed by PCR (Fig. 3). exoC primers reached a band of amplification of 250 bp when using A. brasilense, A. halopraeferens, A. amazonense, A. irakense and A. lipoferum genomic DNA, but no amplified product was detect using A. doebereinerae DNA. Finally, we detected a band of 300 bp in all the samples when primers for rmlD amplification were used. All bands were sequenced and showed 98–99% identity to exoC or rmlD genes from Sp7 (data not shown). Primers designed to amplify exoB gene reached different pattern of amplified bands in all tested Azospirillum strains. A brasilense, A. irakense, A. amazonense and A. doebereinerae showed two bands of amplification. Double amplification was expected since it has been published that A. brasilense has two exoB genes, exoB1 located on the 90-MDa plasmid and exoB2 on the chromosome [8,20]. In order to check if double amplification was consequence of gene duplication, all bands were excised and sequenced. The DNA sequence of the PCR bands, amplified using exoB primers, did not show identity to exoB gene from Sp7. Occurrence of phbC gene was already published by Kadouri et al. [10]. Authors showed that this gene is present in all the azospirilla analyzed in this work.

Figure 3.

PCR using primers for amplification of internal fragments of exoB, exoC and rmlD genes. Line 1, 100 bp DNA ladder (Promega); line 2, blank; line 3, A. halopraeferens; line 4, A. amazonense; line 5, A. irakense; line 6, A. lipoferum; line 7, A. doebereinerae and line 8, A. brasilense.

3.8Occurrence of exoB and exoC genes in other genus

ExoB and ExoC homology search and sequence retrieval were performed by searching the public database (GenBank non-redundant section) with BLAST-P and PSI-BLAST algorithms [21] through the National Center for Biotechnology Information (NCBI; BLAST analysis of ExoB1 (Q59083) and ExoC (P45632) showed low identity with other proteins. Proteins with 58–50% identities to ExoB1 and ExoC were found in member of Δ-proteobacteria, Rhizobiales, Sphingomonadales, and in the close related microorganisms Magnetospirillum magnetotacticum and Rhodospirillum rubrum. A. brasilense exoC gene was found after transfer of a library of Sp7 into an induced S. meliloti exoC mutant [9]. However, A. brasilense ExoC exhibit only 32% identity with the phosphomannomutase of S. meliloti.

Results of BLAST analyses of RmlD and PhbC of A. brasilense have been already published by Jofré et al. [11] and Kadouri et al. [10], respectively. The authors reported identity of RmlD and PhbC to Rhizobiales and R. rubrum, Rhizobiales and Ralstonia eutropha, respectively.


In the present study we showed that during the exponential growth phase Sp7 cells produced high amount of EPS, that was subsequently consumed during stationary and death growth phase (Fig. 2). Also Kefalogianni and Aggelis [22] suggested the degradation of EPS on Azospirillum spp. However, in this work we demonstrate that EPS is not only degraded but also is used a carbon source for growth. It was found that Sp7 support growth using its own produced EPS or from the mutants as a carbon source. These results show the dynamic phenomenon of EPS production, degradation and metabolization during batch growth of A. brasilense.

During exponential growth phase, glucose was the dominant sugar in Sp7 EPS. However, during stationary and death growth phases, arabinose became the dominant one (Table 2). The synthesis of glucose-rich EPS in A. brasilense Sp7 – Cd has been previously reported by Burdman et al. [7] and Fischer et al. [23]. The synthesis of a new EPS, composed of approximately 90% galactose and only traces of other sugars, by A. brasilense growing under saline stress condition, was reported by Fischer et al. [23]. The authors also observed that root exudates induce changes in the EPS composition, with an increase in the amount of arabinose and a decrease in glucose. All those reports focus on point time studies, but we performed a course time study that showed that the composition of the EPS changes during cell growth phase. In addition, to the best of our knowledge this is the first report about growth phase-dependent production of an arabinose-rich EPS by A. brasilense. The production of this EPS occurred during stationary and death growth cell phase of the wild type strain, but it was the dominant EPS during all growth phases of the mutant strains.

Interestingly, Sp7 was unable to grow on l-arabinose monosaccharide as a carbon source, but it grew on the arabinose-rich EPS. Van Bastelaere et al. [24] reported the presence of a SBP (Sugar Binding Protein) encoding gene, sbpA, induced by l-arabinose, d-fucose and d-galactose, suggesting the uptake of l-arabinose in A. brasilense. This gene is part of a large operon that includes ABC-transporters. Upstream to this operon a gene involved in l-arabinose degradation was also found [25]. This genomic organization suggests the possibility of transport of larger macromolecules as reported for alginate (27 kDa) by a pit-dependent ABC transporter (super-channel), which functions like a funnel of macromolecules in Sphingomonas sp. A1 [26]. ABC transporters are the largest and most highly conserved super family of proteins found. They can transport from ions to larger polypeptides and polysaccharides heading by ATP hydrolysis [27]. The model of macromolecules import super-channel proposes the internalization of the polysaccharide by a pit, its binding to substrate binding proteins, delivery through ABC transporter and depolymerization into the cytoplasm [26]. We suggest that A. brasilense might possess a mechanism for internalization of rich-arabinose polysaccharides accompanied by depolymerization of the β-linker and release of monosaccharides into the cytoplasm or by the production of a cell bound glycosidase as was reported for Lactobacillus rhamnosus[28]. The above possibilities are under current investigation.

It has been previously reported that cells of A. brasilense, growing at high C:N ratio, tends to aggregate and flocculate [6] and it was shown that the amount of arabinose present in the EPS correlates with the extent of cell aggregation of different Azospirillum strains [7]. In agreement with those reports our results showed that EPS arabinose content and cell aggregation ability correlate. In addition, it was shown that the rate of arabinose-rich EPS production and cell aggregation were influenced by the energetic status or growth phase of the cells. During exponential phase rich-energy compounds like glucose are stored as biopolymers like EPS. During stationary phase the demand of energy for maintenance purposes decrease, but the depletion of available energy from the medium of growth forces the cells to use their own resources of energy. The oxidative l-arabinose metabolism determines the formation of α-ketoglutarate with the concomitant production of ATP by citric acid cycle and oxidative phosphorylation [29]. To keep working citric acid cycle, acetyl-CoA must be generated.

Acetyl-CoA is the main precursor of PHB synthesis. PHB is synthesized as intracellular energy and carbon storage materials [30]. Here, we showed that the phbC mutant degraded EPS slower than the wild-type strain. Probably, the cells maintain the storage of energy and carbon as EPS. The exoB, exoC and LPS mutants also degraded EPS at slower rate than the wild type. Probably, the storage of energy into PHB is bio-energetically less expensive that into polysaccharides. It has been suggested that the accumulation, degradation and utilization of PHB by bacteria under stress is a mechanism that favors their establishment, proliferation, survival and competitiveness in carbon and energy limiting source environments [31]. Kadouri et al. [32] showed that the phbC mutant exhibits lower survival than the wild-type strain in carrier materials used for soil inoculants. Both PHB accumulation and aggregation could constitute a protected model of growth that allows survival in a hostile environment. The aggregate structure contains channels in which nutrients can circulate and probably can protect the cells from oxidative burst produced during cell death. Kefalogianni and Aggelis [22] found that during EPS degradation by Azospirillum spp. there is a considerable nitrogen fixation, as measured by acetylene reduction activity, and aggregation of cells. The authors proposed that aggregation is probably related to nitrogenase protection from oxygen. The production of a rich-arabinose EPS might have a role in cell aggregation, attachment and colonization processes. Burdman et al. [7] showed that mutants defective in aggregation and reduced root attachment produced arabinose-free EPSs. It has also been reported that the polar flagellum, the outer membrane protein MOMP and cell-surface lectins of A. brasilense function as adhesion factors that are involved in cell aggregation [3]. In a screening for cell-surface lectins it was found that the only sugar capable of completely inhibition of agglutination was arabinose [3]. This and our findings provide evidences of the role of arabinose present in EPS in cell aggregation of A. brasilense. The bacterial cells adhere and attach to roots, plant cell suspensions, abiotic surfaces and to themselves within bacterial aggregates and flocs. The attachment of the bacteria to plant roots is a requirement for the establishment of the bacterial-root association. Cell aggregation could increase survival of Azospirillum cells under diverse stress conditions [3]. This phenomenon may also be important during root colonization where cell aggregation is commonly observed. Advances in the understanding of the aggregation process will also broaden our knowledge of the interactions that lead to the adhesion and colonization of the plant roots by the bacteria.

Differences were observed between Sp7 and phbC mutant cells under SEM. Sp7 cells stayed individually and phbC mutant cells sticky among them. In addition, the glucose-rich EPS produced by Sp7 gave the impression of being shorter needle-like filaments. The arabinose-rich EPS produced by phbC mutant appeared like large needle-like filaments. The production of sticky cells was expected since the arabinose-rich EPS producer exhibited higher cell aggregation capability. These results showed the relationship between composition and mechanical properties of EPS. If the arabinose-rich EPS is stickier that the glucose-rich EPS, an Azospirillum strain able to produce an arabinose-rich EPS will have an advantage when attaching to a root plant. In accordance with our thought, Burdman et al. [7] showed that A. brasilense Sp72002, a mutant defective in aggregation and reduced root attachment, produced an arabinose-free EPS. In addition, the LPS mutant showed impaired attachment to maize roots and diminished maize root colonization [11] and produced glucose-rich EPS.

PCR results suggested that A. brasilense, A. halopraeferens, A. amazonense, A. irakense, A. doebereinerae and A. lipoferum have similar exoC genes for the EPS synthesis and similar rmlD gene for rhamnose synthesis. The occurrence of similar exoB gene on the genus Azospirillum seems to be uncertain. It was reported that the exoB1 gene is located in the recently sequenced plasmid p90 of A. brasilense[8,20]. The G + C content of the p90 and the chromosome are 65.9% and 69–70%, respectively [33]. However, the G + C content of exoB1 is 71%, in accordance with a chromosomally location or consequence of a horizontal transfer. Recently, Vanblue et al. [33] showed that exoB gene is only chromosomally located. Thus, the existence of exoB in Azospirillum needs further clarification. The low degrees of identity of exoB1 with other known proteins suggest this is a unique gene. Finally, BLAST analysis showed that ExoC, PhbC [10] and RmlD [11] related proteins are also present in non-azospirilla related microorganisms.


We thank J. Vanderleyden, E. Jofré and G. Mori for providing us the Azospirillum mutants, S. Wolf and Y. Weiss for their help with HPLC experiments and we also thank S. Burdman for his valuable suggestions during writing. A Lady Davis Trust fellowship supported the work of SCS at The Hebrew University of Jerusalem.