Abstract

1 Introduction

Successful introduction of effective biological control agents in agriculture requires evidence of establishment, proliferation and activity of the inoculants under in situ conditions [1]. Pseudomonas spp. comprise a group of root-colonizing bacteria, which are under consideration for possible utilization as seed inoculants in biological control of plant pathogenic fungi [2]. Evidence for establishment of significant populations of these bacteria on the plant root surfaces is sparse, although many studies have demonstrated their occurrence in rhizosphere soil surrounding the roots [3–9]. Information on the metabolic activity and growth of root-attached Pseudomonas spp. is also limited.

In situ localization of Pseudomonas spp. on barley root surfaces using fluorescence-labelled antibodies targeting specific lipopolysaccharide antigens and detection by confocal laser scanning microscopy (CLSM) has shown a distribution of the bacteria as both single cells and microcolonies (irregular aggregates or strings of cells) between the epidermal root cells [10]. Similarly, the fluorescence in situ hybridization (FISH) technique using 16S rRNA-targeting oligonucleotide probes has become an attractive tool for localization of microorganisms in soil and plant root environments [1,11–14]. Recently, the FISH technique has further been used to distinguish active bacteria in soil [12]. In that study significant fluorescence intensity was observed only in cells with relatively high levels of rRNA (ribosome number) supporting protein synthesis [15], whereas weak or no fluorescence signal was found in starving or stressed bacteria with low ribosome numbers [16].

Our aim in this study was to study the early root colonization of both inoculant and native soil bacteria on sugar beet seedlings in natural soil. The biocontrol strain Pseudomonas fluorescens DR54 [2] was pre-inoculated on the seedlings and root specimens were retrieved at intervals during 3 weeks of seedling development. A strain-specific, fluorescence-labelled antibody was used to localize inoculant DR54 cells on the roots and the general (Eubacteria) rRNA-targeting oligonucleotide probe EUB338 [17] was used to detect activity (ribosomal content) in both the inoculant and native bacteria. When the inoculant Pseudomonas cells had been localized by immunofluorescence, the FISH protocol could thus be used to detect whether the cells were active or not.

2 Materials and methods

2.1 Bacterial strain, culture medium and growth of sugar beet plants

Sugar beet seeds were pregerminated on moist filter paper at 15°C for 5 days, to form roots approx. 5–10 mm in length. P. fluorescens strain DR54 [9] was grown in citrate minimal medium [18] on an orbital shaker (225 rpm) at 15°C overnight until mid-exponential phase (OD 1.7). The bacteria were washed and centrifuged twice (6000×g for 7 min) and finally suspended in 0.9% NaCl (2.5×10⁹ cells ml⁻¹) before coating the young roots. The seedlings (approx. 0.5 cm in length) were coated in the bacterial suspension for 30 min under gentle shaking. Roots from three seedlings were immediately excised and placed in fixation solution (4% paraformaldehyde in phosphate-buffered saline, PBS; pH 7.2) for 4 h. The other seedlings (three for each day of harvesting) were placed under 1 cm of natural soil (Højbakkegård, Tåstrup, Denmark; 15% water w/w) in pots sealed in plastic bags and incubated at 15°C under 12-h cycles of light and dark. Root samples were retrieved after 1, 2, 3, 4, 7, 13, and 20 days of incubation.

2.1 Labelling with antibodies and ribosomal oligonucleotide probe

After excision from the seedlings the roots were washed in sterile PBS (pH 7.2) followed by fixation in 4% paraformaldehyde in PBS for 4 h. The roots were first stained with fluorescein-conjugated antibody against P. fluorescens DR54 and then with CY5-conjugated Eubacteria probe EUB338 [17]. For the dual staining protocol a modification based on several previous reports [1,10,11,14,17,19] was used. The polyclonal antibody against P. fluorescens DR54 was prepared by immunizing rabbits (strain Ssc:CPH) with lipopolysaccharide antigens prepared after proteinase K digestion [20]. Immunization, determination of specificity and removal of weak cross-reactions were performed as described by Hansen et al. [10]. Antibody staining was also carried out according to [10] except that the primary antibody treatment was prolonged from 2 h to 16 h.

After the last washing step in antibody staining, the roots were washed in PBS and transferred to microtubes containing 0.5 ml preheated (50°C) hybridization buffer (0.9 M NaCl, 0.01% SDS, 5 mM EDTA, 20 mM Tris–HCl pH 7.0). A total of 180 ng EUB338 was added and the tubes were incubated at 50°C for 16 h. After hybridization the samples were washed three times in preheated hybridization solution at 50°C for 30 min. Finally, the roots were again washed in PBS followed by sequential transfer through 19.75%, 38.5%, 58.25% and 77% glycerol solutions. Prior to microscopy the roots were mounted in DABCO solution (0.23 g ml⁻¹ DABCO, 10 mM Tris–HCl (pH 8.0), 0.77% glycerol; Sigma) on a slide. A major advantage of this protocol was that only one fixation step with paraformaldehyde was needed before staining.

2.3 Confocal laser scanning microscopy

A confocal laser scanning microscope (TCS4d, Leica Laser Technik GmbH, Heidelberg, Germany) was used to analyze the root specimens [10]. For all sampling days we analyzed three roots at both the base (approx. 3–5 mm below the seed) and the tip (approx. 1 mm behind the apex). A recording consisted of obtaining two stacks of optical sections: one for fluorescein (DR54) and one for CY5 (Eubacteria). Each stack contained 21 horizontal sections with a distance between consecutive sections of 0.5 mm. In the vertical direction we thus obtained data from 10 μm. Each stack was subsequently combined into one image using the maximum intensity procedure. In addition to the standard recordings we examined 1 mm² of a 7-day-old root in detail. Eighty-seven overlapping recordings were made and subsequently puzzled together on a large image. The two fluorochromes used were FITC (excitation wavelength 488 nm, main beam splitter DD, barrier filter BP530) and CY5 (excitation wavelength 647 nm, main beam splitter TK660, barrier filter RG665).

During the 20 days of incubation, any changes of fluorescence intensity for the antibody and ribosomal probe stainings of strain DR54 were carefully recorded. By subsequent exclusion of the DR54 cells in these images, the native soil bacteria could also be distinguished by their staining with the ribosomal probe. A tentative classification of the native bacterial population was made by recording the numbers in three different size classes according to cell width: small cells (<0.6 μm), medium cells (0.6–1.0 μm) and large cells (>1.0 μm). Enumerations of both P. fluorescens DR54 and the native bacteria were carried out visually using screen images of the CLSM data. Occurrence of both single cells and cells within microcolonies (aggregation of at least three cells in close proximity) was recorded. Soil particles rarely distorted the readings and bacteria could always be distinguished by their distinct morphology. Correction for unspecific staining, which was necessary in the recent FISH application of the EUB338 ribosomal probe to stain active bacteria in bulk soil samples [12], was never required in the root specimens of this study.

3 Results

3.1 Root colonization by active population of P. fluorescens DR54 inoculant

Root-colonizing P. fluorescens DR54 were counted at the root base during the 20-day incubation period (Fig. 1). The results demonstrated that a large number of root-attached units (approx. 500–600 per 10⁴μm² root surface) established initially at the root base but were only maintained for 1–2 days. Hereafter the population density suddenly declined to a lower but constant level (approx. 50–100 per 10⁴μm²).

1

View largeDownload slide

Density of P. fluorescens DR54 inoculant on root surface during development of sugar beet seedlings. Mean and S.E.M. expressed as root-attached units (single cells plus microcolonies). Numbers indicate single cells plus microcolony-forming units.

1

View largeDownload slide

Density of P. fluorescens DR54 inoculant on root surface during development of sugar beet seedlings. Mean and S.E.M. expressed as root-attached units (single cells plus microcolonies). Numbers indicate single cells plus microcolony-forming units.

Analysis of the images of FITC-labelled antibody staining and CY5-labelled ribosomal probe staining of P. fluorescens DR54 allowed the cells to be localized as well as their activity to be monitored. Furthermore, the inoculant bacteria could be located whether occurring as single cells or as microcolony-forming cells on the roots. Intensity of the antibody staining (strong or weak) shown in Table 1 indicated that single cells of P. fluorescens DR54 could occasionally be weakly stained whereas cells occurring in microcolonies were always strongly stained. Since weakly stained cells were included in the population density recorded (Fig. 1), variable staining intensity did not influence the general trend of root colonization by the inoculant over time. When compared by their CY5 fluorescence intensity (strong, weak or absent) as shown in Table 1, the decreasing fluorescence intensity indicated that both single cells and microcolony-forming cells lost their activity over time. The most rapid decrease was observed in single cells, for which a weaker staining was already detected on day 3, followed by a complete absence of staining on day 4 for at least some cells. In contrast, all microcolony-forming cells remained strongly stained during at least the first week of incubation before some microcolony-forming cells became weakly stained after the second week.

Intensities in single cells and in cells within microcolonies are shown. S, strongly stained; W, weakly stained; 0, cells visible by antibody but not by the EUB338 probe; –, no microcolonies observed

Fig. 2 shows three selected sets (A, B and C) of dual images based on antibody (Fig. 2, left) and ribosomal probe (Fig. 2, right) stainings of root specimens from days 4, 7 and 13. The elongated root cells indicate the longitudinal direction of the root. The images of antibody staining demonstrated that different distribution patterns may be observed for P. fluorescens DR54 including occurrence as single cells (thin arrows in Fig. 2A), microcolonies of aggregate-forming cells on the epidermal cells (thick arrow in Fig. 2A), microcolonies of string-forming cells between the epidermal cells (thick arrow in Fig. 2B) and microcolonies of chain-forming cells on a root hair (thick arrow in Fig. 2C). The ribosomal probe images in turn provide an estimate of metabolic activity in the DR54 cells during their colonization. At least until day 4, DR54 cells were all active showing either stronger or weaker hybridization signals both as single cells and when occurring in microcolonies (thin and thick arrows in Fig. 2A). From day 7, however, single cells were only detectable with antibody and not by the ribosomal probe (thick arrow in Fig. 2B), while all cells occurring in microcolonies were active during the whole 20-day incubation period (thick arrow in Fig. 2C). The colonization pattern at the root base was however in strong contrast to that at the root tip, where most single cells remained active throughout the 3-week incubation period (data not shown).

2

View largeDownload slide

Dual images of FITC-conjugated antibody (left) and CY5-conjugated ribosomal EUB338 probe (right) staining of root specimens. A: Four-day-old root specimen with single cells (thin arrows) and microcolony (thick arrow) of P. fluorescens DR54. B: Seven-day-old root specimen with string-forming microcolony of P. fluorescens DR54 stained with antibody (thick arrow) but not with EUB338 (right). C: Thirteen-day-old root specimen with chain-forming microcolony of P. fluorescens DR54. The latter is stained with both antibody and EUB338 (thick arrows). Scale bar: 20 μm.

2

View largeDownload slide

Dual images of FITC-conjugated antibody (left) and CY5-conjugated ribosomal EUB338 probe (right) staining of root specimens. A: Four-day-old root specimen with single cells (thin arrows) and microcolony (thick arrow) of P. fluorescens DR54. B: Seven-day-old root specimen with string-forming microcolony of P. fluorescens DR54 stained with antibody (thick arrow) but not with EUB338 (right). C: Thirteen-day-old root specimen with chain-forming microcolony of P. fluorescens DR54. The latter is stained with both antibody and EUB338 (thick arrows). Scale bar: 20 μm.

3.2 Root colonization by active population of native bacterial population

Active cells of the native bacterial population colonizing the root from the bulk soil could be followed by the ribosomal probe. Fig. 3 shows the total cell counts obtained by ribosomal staining for the three subpopulations of native bacteria: small (Fig. 3A), medium (Fig. 3B) and large (Fig. 3C). None of the three subpopulations attached to or proliferated on the roots during the first 2 days of incubation when the P. fluorescens DR54 population was still abundant (Fig. 1). However, intensive colonization by the native bacteria took place thereafter and lasted until day 7, when the population densities approached maximum levels of approx. 50, 1200 and 100 cells per 10⁴μm² for the small (<0.6 μm), medium (0.6–1.0 μm) and large (>1.0 μm) bacteria, respectively. Hence, there was a common and rapid colonization for all three size classes of native soil bacteria, including an initial attachment from day 2 and a maximum of development reached at day 7. While the large subpopulation of medium-sized cells established a constant density at this time, the results indicated that the two smaller subpopulations of small and large cells had a transient appearance on the root.

3

View largeDownload slide

Density of native soil bacteria on root surfaces during development of sugar beet seedlings. Cells were detected by ribosomal staining and fractionated by size into three subpopulations of small, medium and large cells respectively.

3

View largeDownload slide

Density of native soil bacteria on root surfaces during development of sugar beet seedlings. Cells were detected by ribosomal staining and fractionated by size into three subpopulations of small, medium and large cells respectively.

To study root colonization by the native bacteria in detail, we examined a large area (1 mm²) at the root base. Fig. 4 comprises the combined image of 87 standard recordings based on ribosomal probe staining of a root specimen harvested on day 7. Thus the resolution of this large image is the same as for all other images. Both Fig. 4A and the selected enlargements shown in Fig. 4B,C clearly demonstrate that numerous soil bacteria of different size and morphology have colonized the root base at this time, either as single cells or as cells occurring in microcolonies. Large assemblages of both small-sized (thin arrow in Fig. 4C) and large-sized (thick arrows in Fig. 4C) bacteria occur in string-forming microcolonies along the epidermal root cells. In particular, the dense microcolony development of one large and uniform cell type (thick arrows in Fig. 4C) indicated that this was a single, specific organism showing very intensive root colonization.

4

View largeDownload slide

View largeDownload slide

Combined image covering a large surface area (1 mm², A) of a 7-day-old root specimen, as stained by CY5-conjugated ribosomal probe. B and C are selected computer enlargements from the center of A, verifying the occurrence of microcolonies of small (thin arrow) and large (thick arrow) cells of native soil bacteria. Scale bar: 10 μm. The full image (A) can be downloaded from the internet site: http://www.ecol.kvl.dk/research/genmic_res/gm_result_archive.

4

View largeDownload slide

View largeDownload slide

Combined image covering a large surface area (1 mm², A) of a 7-day-old root specimen, as stained by CY5-conjugated ribosomal probe. B and C are selected computer enlargements from the center of A, verifying the occurrence of microcolonies of small (thin arrow) and large (thick arrow) cells of native soil bacteria. Scale bar: 10 μm. The full image (A) can be downloaded from the internet site: http://www.ecol.kvl.dk/research/genmic_res/gm_result_archive.

4 Discussion

4.1 Root colonization by P. fluorescens DR54 inoculant

The sudden, 10-fold decrease of the P. fluorescens DR54 population density at the root base occurring after 1–2 days could have several reasons. (1) The decrease could be due to the stretching of root tissue occurring soon after root formation, resulting in longer distances between the cells initially colonizing the root. This explanation was supported by the formation of long, string-forming assemblages of bacteria between the epidermal root cells (Figs. 2 and 4); a similar observation was made during early colonization of barley roots by P. fluorescens strain DF57 [10]. (2) The decrease could also be due to detachment of early-colonizing cells from the root surface and into the slime matrix (mucigel) or the rhizosphere soil. Actually, the recovery of P. fluorescens DR54 cells in the rhizosphere soil after 20 days of incubation (data not shown) clearly demonstrated that detachment was indeed taking place. (3) Finally, a sudden weakening or complete absence of antibody staining during the incubation period could also result in the dramatic decrease of cell density at the root base; however, this was considered unlikely since there was no general trend for a loss of staining intensity over time.

After the initial and sudden decrease, P. fluorescens DR54 established a constant population density on the sugar beet root. Only a small fraction of the total population actually developed into microcolony-forming cells, but the strong fluorescence intensity of the ribosomal probe indicated a high activity in these dividing cells. One of the microcolony types was a string of loosely associated cells (Fig. 2B) which occurred either in grooves between the root epidermal cells or on their surface. As mentioned above, this type most likely formed already during root elongation, as was shown in the earlier study of P. fluorescens strain DF57 and AG 1 colonizing barley seedling roots [10]. Another microcolony type of irregular cell aggregates (Fig. 2A) from day 3 was actually the most common on the sugar beet roots. This type was in turn most likely to have developed from single cells proliferating at a later stage, i.e. after the initial phase of root elongation. Finally, the rare microcolony type of tightly associated cells on root hairs (Fig. 2C) was clearly also developed at a relatively late stage of root development.

In contrast to the dividing microcolony-forming cells, a majority of the single cells of P. fluorescens DR54 soon decreased their cellular activity (ribosome numbers) after initial attachment at the root base (Table 1). It was thus evident that a majority of the single cells on the root never progressed to a first cell division and subsequent microcolony formation. A possible explanation could be that the cells showing low or absent hybridization signal with the ribosomal probe entered an inactive or dormant stage, representing a low number of ribosomes per cell. When compared with the constant, overall population density recorded by immunofluorescence staining, the evidence of active and non-active subpopulations recorded by ribosomal staining provides new and direct information on inoculant survival at the single-cell level. The differentiation into a significant subpopulation of inactive single cells at the root base was most likely due to declining nutritional resources over time, e.g. resulting from increasing competition with the native soil bacteria colonizing the root (see below). Another explanation could be that the major exudation zone for soluble organic substrates associated with the root tip became more distant with time. In support of this, it was noticed that most single cells of P. fluorescens DR54 detected by antibody staining at the root tip remained active as judged from the staining with ribosomal probe. This is the first direct documentation by microscopy of active, root-colonizing cells of the inoculant, following the root apex during its progression in the soil matrix.

4.2 Root colonization by native bacterial population

A native population of soil bacteria colonized the root base intensively during the first 7 days of incubation. Interestingly, however, intensive colonization took place only after a delay until day 2, when the P. fluorescens DR54 population had decreased to a lower, but constant density. From this time, the native population increased steadily on the roots until day 7 and the composite images covering a large surface area of the root supported that massive colonization took place at the root base during this period. The initial delay of 1–2 days prior to colonization followed by an intensive colonization phase of 5–6 days may have important implications also for the performance of the P. fluorescens inoculant colonization pattern, survival and efficacy of biological control. Hence, if competition between the inoculant and native bacterial populations were important when substrate resources are declining, this interaction may add an explanation for why a large proportion of the inoculant P. fluorescens cells became inactive with time.

An indication of competitive interaction also between the native soil bacteria was provided by the two small subpopulations of relatively small and large-sized cells, respectively, showing a transient appearance by their decreasing numbers of active cells during extended incubation. The results further indicated that at least one specific organism representing a thick, short rod was predominant mostly during the intensive phase of root colonization and thus may represent a transient appearance on the root. To confirm and further resolve the evidence for successions occurring within the root-colonizing native bacteria, additional studies should be performed using ribosomal probes targeting selected groups of soil bacteria.

Acknowledgements

This study was supported by the Danish Biotechnology Programme, the Danish Strategic Environmental Programme (Center for Effects and Risks of Biotechnology in Plant Production) and Danish Agricultural and Veterinary Research Council Grant 9311819. The excellent technical assistance of Bente Østergård is gratefully acknowledged.

References