INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The rhizosphere is a highly important source of organic materials in soils. It contributes to soil organic matter directly through the release of soil root material and its decomposition and indirectly through root exudation, which stimulates microbial activity and biomass (22). The complex and dynamic interactions among microorganisms, roots, soil, and water in the rhizosphere induce changes in soil physicochemical and structural properties. For example, long-term pasture soils which were permanently and densely rooted had greater aggregate water stability than did their long-term arable counterparts (15, 39). Aggregate stability of ryegrass and alfalfa rhizosphere soil has been shown to increase with root growth, and the stabilizing effect apparently originated from polysaccharide production in the rhizosphere (29). The efficiency with which ryegrass increases the size of soil aggregates has been attributed to the activation and growth of mycorrhizal and saprophytic fungal hyphae (23, 38). Inoculation of wheat with Paenibacillus polymyxa in a silty topsoil resulted in an increase in the amount of root-adhering soil (RAS) per unit of root tissue (RT) dry mass (dm) and enhanced the frequency of the aggregate size class of 0.2 to 2 mm (13). The soil water potential determines (i) the availability of water, oxygen, and substrates for plants and microorganisms, (ii) the survival and activity of microorganisms, and (iii) competition and predation relationships among microorganisms (3, 26, 31, 40). Bacterial exopolysaccharides (EPS) can protect bacteria from various stresses. A Pseudomonas sp. strain increased its EPS production during desiccation (30). The production of EPS possibly enhances water retention in the microbial environment and seems to regulate the diffusion of carbon sources such as glucose (5, 6). In Mediterranean agriculture, hydric stress in soil is a major factor limiting crop production. Therefore, we isolated an EPS-producing bacterium from wheat roots growing in a Moroccan clayey soil and characterized, by a combined microbiological and physical approach, the effects of inoculation of EPS-producing bacteria and of soil water content on colonization of the wheat rhizosphere and on rhizosphere soil aggregation.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Soil. Soil was collected in September from the homogeneous A1 horizon (0 to 20 cm) in western Morocco (Institut National de la Recherche Agronomique experimental station, Settate). It was previously under a wheat-legume crop rotation. Due to the following characteristics, it was classified as a vertisol (Food and Agriculture Organization classification): clay, 55.1%; fine silt (2 to 20 µm), 21.5%; coarse silt (20 to 50 µm), 9.7%; fine sand (50 to 200 µm), 10.2%; coarse sand (>200 µm), 1.7%; total organic carbon, 1.47%; total nitrogen, 0.10%; pH (H₂O) 8.2. As clay minerals, we determined smectite (80%), kaolinite (10%), and chlorite (10%) after clay ultrafractionation (33) and X-ray diffraction analysis. This soil contained 22% carbonates and was saturated with calcium. The soil sample was air dried, sieved to 2 mm, and stored in plastic containers at 4°C (7% soil water content) until use.

Selection of a bacterial strain. Bacteria were isolated from the wheat rhizosphere 3 weeks after wheat cultivation in plastic pots (5 by 18 cm) containing a soil sampled in Settate with an average soil water content of 24%. This soil sample was not air dried. Bacteria were isolated from root macerates, RAS, and nonrhizosphere soil. Suspensions were serially diluted in 0.85% KCl and plated on agar containing the ingredients of modified Weaver's medium (yeast extract, 0.1 g liter¹; supersalts solution, 50 ml liter¹; phosphate buffer, 15 ml liter¹) (17) and also sucrose, fructose, sorbitol, glucose, or mannitol (20 g liter¹). Rhizoplane strain NAS206 was chosen for its extensive EPS production on the sucrose-supplemented medium. This strain was phenotypically characterized by using the Biolog system (Biolog Inc., Hayward, Calif.). Inoculated Biolog GN microplates were incubated at 30°C and examined for 4 and 24 h, and bacteria were identified by means of the substrate utilization profile by using Biolog Microlog software. Strain NAS206 was also identified by using Biotype-100 strips (bioMérieux), which is more suitable for the identification of members of the family Enterobacteriaceae. Data were analyzed by P. A. D. Grimont (Institut Pasteur, Paris, France), who used the Recognizer program. Amplified ribosomal DNA (rDNA) restriction analysis of strain NAS206 and the type strains of Pantoea dispersa (LMG 2603^T), P. agglomerans (E20^T; kindly provided by P.D.A. Grimont), and Rahnella aquatilis (ATCC 33071^T) was performed to verify the phenotypic identification. The total 16S rDNAs of the four strains were amplified by PCR and digested with 13 restriction endonucleases (AluI, CfoI, DdeI, HaeIII, HinfI, TaqI, MspI, RasI, NciI, Sau961, ScrfI, NdeII, and Tru91) (21, 42).

Inoculation with strain NAS206 and wheat growth conditions. Wheat seeds (Triticum durum L. cv. Acsad 65) were surface sterilized with saturated calcium hypochlorite (2 h) and 10% H₂O₂ (30 min) under a partial vacuum. The seeds were rinsed in sterile water and incubated on nutrient agar plates (Difco) at room temperature to check for sterility. Prior to inoculation, strain NAS206 was grown in Luria-Bertani broth (34) at 30°C for 24 h. Cells were harvested by centrifugation (12,000 × g, 10 min), washed twice (0.85% KCl), and then resuspended in sterile distilled water to a density of 10⁸ bacteria ml¹. Sterilized wheat seeds were coated with 4 g of peat containing 1 ml of a bacterial inoculum (giving 4 × 10⁷ bacteria per seed) and 4 ml of EPS from strain NAS206 (35%, wt/vol). The EPS was extracted as described by Hebbar et al. (16), from a solid medium (24) containing 2% sucrose which was inoculated with strain NAS206 and incubated at 30°C for 16 h and then at 4°C for 24 h. As the control, seeds were coated with the peat mixed with EPS, but with no bacterial inoculum.

Colonization of soil and roots by strain NAS206. Colonization of wheat roots, rhizosphere, and nonrhizosphere soil by strain NAS206 was studied with 20-day-old plantlets (five replicates per treatment). The entire soil-root system was taken from the pot and agitated gently, giving the bulk soil fraction. The roots were washed by being dipped in sterile water to separate RAS from root tissue RT. To localize strain NAS206 at different wheat root parts, the washed root system was aseptically dissected into three segments: the first 5 cm below the seed (basal region), 4 cm of the middle of the root system, and 2 cm at the root tip. The different root macerates, rhizosphere soil, and bulk soil were serially diluted in 0.85% KCl and plated on petri dishes containing nutrient agar. After incubation at 30°C for 24 h, yellowish pigmented colonies (similar to strain NAS206) and total microflora CFU were counted from appropriate dilutions. To verify that the yellow colonies represented the strain NAS206 inoculum, 24 colonies were randomly selected and checked for their genetic fingerprint by PCR with enterobacterial repetitive intergenic consensus (ERIC) sequences as primers (19). The protocol was adapted from that of de Bruijn (8, 35) and recently described in detail (10).

Characteristics of RAS. Six plantlets per treatment were randomly chosen 20 days after sowing. Roots with adhering soil were carefully separated from bulk soil by gentle mechanical agitation (Agitest; Bioblock) for 1 min. RAS was removed from RT by washing in distilled water. Soil dm and root fraction dm were measured after 24 h at 105°C, and the RAS/RT ratio was calculated. Water stability of RAS was determined by using the wet-sieving method (2). Each root system with adhering soil was placed onto a sieve (height, 70 mm; diameter, 60 mm; mesh size, 200 µm), immersed in 200 ml of distilled water, and oscillated. The horizontal oscillations applied for 1 h were sinusoidal (2-cm amplitude, 98 oscillations min¹). Amounts of water-stable (>200 µm) aggregates were calculated by substracting coarse sand (1.7%) and root fragments remaining on the sieve. After 1 h of disaggregation, Na-Amberlite resin IR-120 (500 µm) was added to the soil suspension (<200 µm). To remove clay particles from microaggregates, the suspensions were agitated for 16 h with an end-over-end shaker at 40 rpm. Sand, silt, and clay particles were separated by sieving and sedimentation by using the Robinson pipette method (32).

3

1

Statistics. Analyses of variance were performed by using the Statgraphics software program (version 5.0; STSC software products). The least-significant-difference test (Newman-Keuls test) was used for multiple-range analyses.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Origin, identification, and EPS production of strain NAS206. Strain NAS206 was isolated from the root surface of wheat which was cultivated in the Settate vertisol studied. On sucrose-supplemented Weaver's medium, strain NAS206 produced an EPS which was composed of galactose and glucose as monosaccharides and of succinate and pyruvate subunits (17a). Strain NAS206 was identified as P. dispersa by using Biolog microtiter plates, and as P. agglomerans by using the Biotype-100 system. The 13 restriction endonucleases tested revealed a strong polymorphism of the amplified rDNA restriction analysis profiles of NAS206, P. dispersa LMG2603^T, P. agglomerans E20^T, and R. aquatilis ATCC 33071^T. Strain NAS206 was differentiated from R. aquatilis ATCC 33071^T by 32 sites and from P. dispersa LMG2603^T by 12 sites but from P. agglomerans E20^T only 2 sites (obtained by using DdeI). This indicates that the species closest to strain NAS206 is P. agglomerans, which is in agreement with the phenotypic identification by Biotype-100. It was concluded that strain NAS206 belongs to the species P. agglomerans.

Soil hydric conditions. Under each hydric condition tested, the water content of bulk soil in noninoculated and inoculated treatments was statistically the same: 35% ± 0.7%, 24% ± 0.9%, and 20% ± 0.6% for the nearly water-saturated, drained, and dry soil variants, respectively.

Soil, rhizosphere, and root colonization by P. agglomerans NAS206. In the noninoculated treatment, after 20 days of wheat growth, the size of the spontaneous P. agglomerans population was about 10³ CFU g of dm¹ in bulk soil and 10⁵ CFU g of dm¹ on the root surface (rhizoplane), representing, respectively, less than 0.003 and 3% of the total microflora plate counts. The population size of P. agglomerans in bulk soil was not affected by inoculation of strain NAS206, whereas in the RAS fraction, strain NAS206 appeared to be strongly proliferating and represented about 50% of the total microflora in the dry variant, where its proliferation was most intense (Table 1). The average rhizosphere effect on the P. agglomerans population (number of bacteria in the rhizosphere/number of bacteria in the bulk soil) was 2,540 ± 685. To validate the counts of strain NAS206, 24 yellow and mucoid colonies were randomly selected and submitted to PCR analysis. They provided ERIC-PCR profiles identical to that of strain NAS206 (data not shown).

1

1

Effects of soil water content and inoculation on plant growth. After 20 days of wheat growth, there was a slight positive effect (not significant) of increasing soil water content on wheat shoot production (Fig. 1) and on shoot/root ratios, which increased from 1.6 to 2.0. Root growth seemed to be inhibited at a soil water content of 35% (Fig. 1). For each soil water content tested, there was no significant effect of inoculation on plant growth parameters. However, there was a significantly positive overall effect of inoculation on shoot growth, as shown by variance analysis.

View larger version (47K): [in this window] [in a new window]   FIG. 1.   Effect of inoculation with P. agglomerans NAS206 on wheat at various levels of soil water content. Values with different letters are significantly different at P < 0.05 by the Newman-Keuls test. Error bars show standard errors of mean values.

Effect of soil water content and inoculation on rhizosphere soil aggregation. To evaluate the quantitative effect of inoculation and soil water content on soil aggregation, the biomasses of RAS and RT were measured for each wheat plantlet. In control treatments, the RAS/RT ratio was only increased at a soil water content of 35%, compared to a soil water content of 20% (Fig. 2A). There was a positive effect of inoculation on the RAS/RT ratio at each soil water content tested (Fig. 2A). At a 24% soil water content (matric potential of 0.20 MPa), 140 mg of soil was attached to 1 mg of root tissue in the inoculated treatment, compared to the control value of 90 mg of attached soil. Maximum water stability of rhizosphere aggregates was measured for the drained variant of both control and inoculated treatments (24% soil water content, Fig. 2B). Under these conditions, the effect of inoculation was significantly positive, and there was an insignificant trend toward more water-stable aggregates due to inoculation in the other two soil water regimens (Fig. 2B).

View larger version (35K): [in this window] [in a new window]   FIG. 2.   Cross effects of inoculation with P. agglomerans NAS206 and soil water content on the RAS/RT ratio (A) and the water stability of RAS (B). Values with different letters are significantly different at P < 0.05 by the Newman-Keuls test. Error bars show standard errors of mean values.

Effect of soil water content and inoculation on soil porosity. Pore and throat volume distribution curves were always trimodal, with a main mercury macropore volume (V) occurring at throat radii between 3 and 50 µm and two mostly minor mercury micropore volumes at throat radii between 0.05 and 3 µm (v₁) and at throat radii of <0.04 µm (v₂) (Fig. 3A). Such a distribution could possibly be attributed to pores located between silt-clay assemblages for V, i.e., with throat widths ranging between 5 and 20 µm, as measured on the whole set of pores which were observed by transmission electron microscopy (TEM) (Fig. 4A). Intermicroaggregate porosity (Fig. 4B and C) and clay domain porosity (Fig. 4D), respectively, corresponded to v₁ and v₂. For these <2-mm aggregates, this could be envisaged as filling of large pores, followed by filling of small pores of sizes similar to those of clay microcracks (clay fabric porosity).

View larger version (25K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   FIG. 3.   (A) Effect of bacterial inoculation and soil hydric treatment on the mercury pore volume of RAS and bulk soil aggregates. (B) Effect of inoculation with P. agglomerans NAS206 and soil hydric treatment on the cumulative mercury pore volume of RAS and bulk soil aggregates.

View larger version (97K): [in this window] [in a new window]   FIG. 4.   TEM of the wheat rhizosphere showing the ultrastructure of RAS at various magnifications. Panels: A, whole set of pores; B and C, intermicroaggregate porosity; D, clay domain porosity.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Wheat rhizosphere colonization by P. agglomerans NAS206. Inoculation of wheat seeds with strain NAS206 led to pronounced colonization of the rhizoplane of wheat and of rhizosphere soil by this strain. Expressed as a percentage of the total microflora, colonization by this strain was significantly more intense in the basal root section than in the other root zones tested (Table 2). The ability of this strain to form cell aggregates (symplasmata) at the rhizoplane could explain its heterogeneous distribution along the wheat roots (data not shown). Passive transport by root elongation might also be a mechanism of bacterial wheat root colonization. In the absence of water movement, root elongation has been reported to be the major factor in bacterial colonization of wheat by Pseudomonas fluorescens and Bacillus subtilis (9, 18). The relatively low colonization of root tips by strain NAS206, particularly at high water content, may be attributed to the nature of root exudates. It has been shown that low-molecular-weight compounds are rapidly expelled from the root tip into the surrounding soil, explaining why bacterial population densities are always low at the root tip (25).

Effect of soil water content and inoculation on rhizosphere soil aggregation. The enhanced population size of inoculated strain NAS206 in the wheat rhizosphere was associated with an increase in the RAS/RT ratio (Fig. 2A). This is probably due to an aggregating effect of the EPS produced by the inoculated bacteria. EPS can increase soil-root adhesion and/or the mechanical stability of the rhizosphere soil, as previously demonstrated for clay minerals by polysaccharide adsorption (4).

Effect of soil water content and inoculation on soil porosity. V, which corresponds to a rapid rise in the mercury injection curve, is interpreted as a connected percolation pore network (12, 37). Such percolation pore networks with neck radii of 5 to 20 µm were directly observed in thin sections studied by optical microscopy (results not shown). The V of either RAS or bulk aggregate soil of control treatments progressively increased from the nearly water-saturated soil to the dry soil, whereas v₁ and v₂ tended to decrease (Fig. 3). This is attributed to shrinkage and associated growth of microcrack networks (Fig. 4), as reported elsewhere (36). The air-drying and vacuum-drying methods applied for mercury porosimetry could induce a strong increase in these processes.