TEXT

Top Abstract Text References

Germinating seeds and growing plants influence the activities of soil microorganisms in the adjoining volumes of soil known as the spermosphere and the rhizosphere, respectively (24). Conversely, microorganisms in these settings condition the seeds and plants in a number of ways. Some of the microorganisms (e.g., the so-called plant-growth-promoting rhizobacteria) may enhance plant health and productivity by synthesizing phytohormones, increasing the local availability of nutrients, facilitating the uptake of nutrients by the plants, decreasing heavy metal toxicity in the plants, antagonizing plant pathogens, and inducing systemic resistance in the plants to pathogens (5, 8, 9). Detrimental effects are produced by other organisms, such as the so-called deleterious rhizosphere microorganisms, and these effects include release of toxic products of microbial metabolism, alteration of nutrient cycling, impairment of uptake of nutrients, competition for nutrients, and retardation of root growth (28). Among the factors involved in plant-microbe interactions, as well as in microbe-microbe interactions, in the rhizosphere, siderophores and hydrocyanic acid (HCN) have received special attention. Siderophores are high-affinity Fe³⁺ chelators that are synthesized and released extracellularly under iron limitation conditions, where they make otherwise inaccessible supplies of insoluble iron available to organisms with specific membrane-bound siderophore receptors (14, 17). Such receptors are present in the microorganisms that produce the corresponding siderophores but can also be found in plants. The iron nutrition of these plants is thus enhanced. In addition, since plant-growth-promoting rhizobacteria produce siderophores with higher Fe³⁺ affinity than the siderophores produced by deleterious rhizosphere microorganisms, the latter microorganisms are outcompeted in their quest for iron. HCN is released as product of secondary metabolism by several microorganisms and affects sensitive organisms by inhibiting the synthesis of ATP mediated by cytochrome oxidase (22). Therefore, depending on the target organisms, HCN-producing microorganisms are regarded as harmful when they impair plant health and beneficial when they suppress unwanted components of the microbial community (23, 29). The significance of siderophore and HCN generation in rhizosphere management and engineering has been studied and reviewed extensively (11, 13, 16, 21, 26, 36), but there is relatively little information on comparative population dynamics and interactions of producing and nonproducing microorganisms during the early stages of root colonization. These issues were specifically addressed in the present work.

Experimental procedures. The following strains were selected from a collection of isolates of Pseudomonas spp. established previously (7): CC13 and CC19 from nonrhizosphere soil close to an isolated plant of mahaleb cherry; CC148 from the rhizosphere of broccoli rab; CC209 from the rhizosphere of leaf beet; CC219/2 from the rhizosphere of artichoke; and CC295 from nonrhizosphere soil located between four blocks containing artichoke, cauliflower, leaf beet, and onion. These strains were identified as Pseudomonas aeruginosa (CC19), Pseudomonas aureofaciens (CC295), Pseudomonas fluorescens biovar 1 (CC13, CC209, and CC219/2), and Pseudomonas putida biovar A (CC148) on the basis of fatty acid methyl ester profiles as described by Stead (33). Table 1 shows the identifying characteristics that were used routinely in the present work to differentiate the strains with respect to production of siderophores with the chromeazurol S test (30), with respect to production of HCN with the test devised by Castric and Castric (6), and with respect to utilization of maltose, sorbitol, or methyl -D-glucopyranoside at a concentration of 2 mg ml¹ as a sole C source for growth in the synthetic medium of Ayers et al. (3) solidified with 15 g of agar per liter. The strains were maintained on 2% (wt/vol) glycerol nutrient agar at 4°C.

1

1

Single-strain inoculations. The average total number of bacteria was 2.1 × 10⁵ ± 0.19 × 10⁵ CFU cm of rootlet¹ 2 h after inoculation of either cucumber or spinach seedlings with any strain (Fig. 1). Most of each count was contributed by fraction L (9.1 × 10⁴ ± 1.27 × 10⁴ CFU cm¹), followed by fractions R (6.8 × 10⁴ ± 0.75 × 10⁴ CFU cm¹) and F (4.8 × 10⁴ ± 0.34 × 10⁴ CFU cm¹). Later counts did not differ significantly (P = 0.872) from the initial values for any of the strains unable to synthesize siderophores except CC295 on spinach, where the values for fractions L and R were lower (P < 0.05) and the value for fraction F was higher (P < 0.05) at 48 h (Fig. 1 A through D). In contrast, the values for all fractions of strains that produce high levels of siderophores increased 2.5- to 3.8-fold on cucumber and 3.5- to 5.0-fold on spinach during the experiments (Fig. 1I through L). With strains that produce moderate levels of siderophores, the counts for fraction L and the counts for fractions R and F exhibited the same trends as the counts obtained for strains that produce high levels and for strains that do not produce siderophores, respectively (Fig. 1E through H). The length of cucumber rootlets increased significantly from 3.3 ± 0.37 to 4.57 ± 0.59 cm at 48 h and from 5.74 ± 0.52 to 8.65 ± 0.69 cm at 96 h after inoculation with strains that produce moderate or high levels of siderophores, independent (P = 0.747) of production of HCN (Fig. 1E, F, I, and J). None of the strains affected elongation of spinach rootlets (P = 0.901), but dark necrotic strips of tissue were visible in the periderm 48 to 96 h after inoculation with HCN producers. The nature of the different responses of cucumber and spinach rootlets to HCN- and siderophore-producing strains was not investigated. When 3-day-old seedlings were dipped in SDW and then grown for 4 days on filter paper saturated with 10⁴ M FeCl₃, the final length of cucumber rootlets increased significantly (P < 0.05) from 6.09 ± 0.58 to 8.60 ± 0.44 cm, but elongation of spinach rootlets was not affected (P = 0.844), suggesting that cucumber, but not spinach, was iron limited under the conditions used. Additional observations were made with seedlings that were raised initially for 3 days as usual and then dipped in SDW and kept for 4 additional days on SDW-impregnated filter paper in two opposite compartments of an X dish in which one of the HCN-producing strains was growing on 2% peptone agar in the other two compartments. In this setting, mild necrotic symptoms similar to those described above were seen on the rootlets of spinach but not on the rootlets of cucumber, supporting the hypothesis that there was etiologic involvement of a gaseous bacterial product.

View larger version (37K): [in this window] [in a new window]   FIG. 1.   Bacterial populations and rootlet elongation after inoculation of cucumber and spinach seedlings with the following bacterial strains: CC13, CC19, CC148, CC209, CC219/2, and CC295. Separate counts are given for total bacteria () and for the bacteria that were considered loosely adsorbed (), reversibly adherent (), and firmly anchored () to the rootlet surface. The length of inoculated rootlets () is shown along with the length of the controls (). Each value is the mean based on two independent experiments performed with triplicate sets of 10 seedlings. Distinguishing characteristics of bacterial strains are shown in Table 1.

Two-strain inoculations. The total and fractional counts for all strains in each pair up to 48 h after inoculation did not differ significantly (P = 0.866) from the counts obtained during the same period when the same strains were inoculated separately (Fig. 2). However, significant differences compared with the single-strain experiments emerged later. When a siderophore producer was coinoculated with a nonproducer, the counts for all fractions of the latter declined significantly (P < 0.05) by 1.3 to 1.8 log units between 48 and 96 h (Fig. 2A through D). Consequently, the total populations of these strains decreased by the corresponding values. This was true whether both strains in the inoculum did or did not produce HCN. For pairs in which the strains differed in production of HCN, the values for fractions R and F of the nonproducer declined by 4.5 × 10⁴ ± 0.9 × 10⁴ to 10 × 10⁴ ± 2.2 × 10⁴ and 3.9 × 10⁴ ± 0.55 × 10⁴ to 6.9 × 10⁴ ± 1.2 × 10⁴ CFU cm¹, but the values for fraction L increased by approximately the same amounts (Fig. 2E through H). Therefore, the total populations reached the same levels (P = 0.831) as those in single-strain experiments whether both strains in the pair were HCN producers or nonproducers. For combinations in which a strain positive for production of siderophores or HCN was paired with a strain negative for the same characteristic, the total and fractional populations at 96 h did not differ significantly (P = 0.838) from those recorded at the same time after inoculation of the positive strain alone. Whenever one or both strains in the inoculum produced siderophores, elongation of cucumber rootlets at 48 and 96 h was significantly greater (P < 0.05) than elongation in the controls (Fig. 2A, B, and F). Necrotic symptoms similar to those described above for single-strain inoculations were seen on spinach rootlets beginning 48 h after inoculation with any combination of strains when one or both of the components produced HCN.

View larger version (60K): [in this window] [in a new window]   FIG. 2.   Bacterial populations and rootlet elongation after inoculation of cucumber and spinach seedlings with the following pairs of bacterial strains: CC148 plus CC219/2, CC13 plus CC295, CC295 plus CC219/2, and CC13 plus CC148. Separate counts for each strain in every pair (open and solid symbols) are given for total bacteria ( and ) and for the bacteria that were considered loosely adsorbed ( and ), reversibly adherent ( and ), and firmly anchored ( and ) to the rootlet surface. The length of inoculated rootlets () is shown along with the length of the controls (). Each value is the mean based on two independent experiments performed with triplicate sets of 10 seedlings. Distinguishing characteristics of bacterial strains are shown in Table 1.

Conclusions. It has been reported widely (4, 12, 15, 27, 31, 34) that the relative importance of biotic and abiotic determinants in microbial and plant-microbe interactions in the rhizosphere can be determined more precisely in gnotobiotic systems than in the field, where heterogeneous, mostly ill-defined factors come into play. The findings of the present work, which were obtained under gnotobiotic conditions, combined with previous reports (1, 2, 18, 20, 32, 35), emphasize that proper assessment of the influence of bacterial properties on the outcome of root colonization requires careful consideration of the test plant and of the indicator effects that must be taken into account. This view is supported by several lines of evidence with respect to production of siderophores and HCN. First, significantly larger populations on cucumber and spinach rootlets were invariably generated by bacterial strains that produced siderophores than by strains that did not. With moderate siderophore producers, however, this effect resulted exclusively from an increase in the fraction of total bacteria that were loosely adsorbed to the rootlet, whereas all fractions of strains that produced high levels of siderophores proliferated more irrespective of the tenacity of their association with the surface of the rootlet. Second, bacterial interactions resulting from the coexistence of different strains on the same rootlet were or were not reflected in the total population size of each strain on the rootlet, depending on whether the strains differed in production of siderophores or of HCN. Third, inoculation with siderophore producers resulted in greater elongation of rootlets on cucumber seedlings but not on spinach seedlings. And fourth, neither the health nor the elongation of cucumber rootlets was apparently disturbed by HCN producers that induced necroses on the rootlets of spinach. Verification of these findings in long-term experiments with exposure to field soil variables is under way.