Bacterial Growth and Development in the Rhizosphere of Gnotobiotic Cereal Plants

Transcription

1 Journal of General Microbiology (198 I), 125, Printed in Great Britain 95 Bacterial Growth and Development in the Rhizosphere of Gnotobiotic Cereal Plants By R. A. BENNETT AND J. M. LYNCH* Agricultural Research Council, Letcombe Laboratory, Wantage, Oxon OX1 2 9JT, U.K. (Received 12 November 1980) The growth and development of a Curtobacterium sp., a Mycoplana sp. and a Pseudomonas sp. in the rhizosphere of gnotobiotic wheat, barley and maize plants were studied. When inoculated singly in the wheat or barley rhizosphere, each of the three species grew exponentially during the first h and slowly thereafter. However, in the maize rhizosphere the Curtobacterium sp. showed only a small increase in numbers and the Mycoplana sp. had a long lag phase, a slower growth rate and a smaller final population than when it was inoculated into the wheat or barley rhizosphere. Interactions between these three species in various combinations and in different rhizospheres are described and some of the possible mechanisms involved are discussed. INTRODUCTION Quantitative studies on the growth and development of microbial populations around plant roots, i.e. in the rhizosphere, have been few, and recent reviews of bacteria-root interactions (Bowen, 1980; Elliott et al., 1981; Lynch, 1981; Newman, 1978) have emphasized the need for more. Barber & Lynch (1977) showed how bacterial growth is related to substrate supply from the roots and how this is relevant to energy supply to associated nitrogen-fixing organisms. Newman & Watson (1977) have shown how the results of quantitative growth studies can be used to provide a model of the rhizosphere. Bowen (1979, 1980) suggests that the kinetics of microbial growth around roots can be understood only by studying plants grown in soil, but in our view the physical, chemical and biological complexity of soil makes for unacceptable difficulty in devising reproducible experiments. Accordingly, we now report experiments with plants grown in tubes of sterile sand. This is more realistic than growth in solution as it provides solid surfaces of the kind that are important for the growth of micro-organisms in nature (Berkeley et al., 1980); it also enhances the development of nutrient gradients. On the other hand, sand avoids many of the complications of soil and allows the growth of sterile plants in a readily defined environment. These plants can then be inoculated with one or more known micro-organisms, making a relatively simple system. Investigations of how one plant species interacts initially with a single microbial species and subsequently with several species in a defined system (autecology) are complementary to studies in soil (synecology): both approaches are necessary to understand the rhizosphere. The numbers of bacteria in the rhizosphere can be estimated from direct counts by microscopy or from plate counts. Discrepancies between the two methods are usually large, particularly with soil-grown plants; for example, Rovira et al. (1974) found that for eight plant species direct counts were 10 times greater than plate counts. The main problems with the direct method are that the bacteria form colonies or aggregates unevenly distributed over the root surface (Rovira, 1956; Dart, 1971; Asanuma et al., 1979), and that long periods of observation are required. In contrast, the main problems with the plate method are that the /81/ $ SGM

2 96 It. A. BENNETT AND J. M. LYNCH chosen medium will inevitably favour the growth of some microbial species more than others in a mixed population; only viable cells will be counted and aggregates of cells will appear as single colonies. These factors may account for some of the discrepancies between results from the two methods. Many of these difficulties can be overcome if gnotobiotic plants are used and the growth requirements of the associated microflora are known so that all viable cells can potentially be cultured on agar plates. We have used this technique with bacteria which seem to be important colonists of barley, maize and rape roots. The colony forms of the bacteria were all distinct and so mixed populations could also be estimated by plate counting. METHODS Apparatus and preparation of sterile plants. Plants were grown in glass tubes containing sterile, acid-washed, coarse sand (Fig. 1). The entire apparatus was sterilized by autoclaving at 121 "C for 30 min. Prior to planting, the sand was moistened by flooding the tubes with sterile plant culture solution (PCS) and allowing them to drain freely. The PCS contained (per litre): Ca(N0,),.4H20, g; KNO,, g; KH2P0,, g; MgSO,. 7H20, g; NaNO,, g; ferric EDTA, 3.5 mg; H,BO,,0.57 mg; CuSO,. 5H20, 0.04 mg; KCI, 1-05 mg; MnSO,. 4H20, mg; (NH,),Mo,O,, mg; ZnSO,, 0.22 mg. All the components were autoclaved (121 "C, 15 min) together except for the ferric EDTA which was autoclaved separately and added to the bulk solution when cool. Wheat (Triticum vulgare var. Mardler), barley (Hordeum vulgaris var. Midas) and maize (Zea mays var. LG11) seeds were sterilized (Barber, 1967) and allowed to germinate for 24 h on moist sterile filter paper before planting in the tubes. Germinating sterile seeds were planted to a depth of 2 cm and the tubes were incubated in a growth cabinet with a 16 h light/8 h dark cycle and a constant temperature of 18 OC. Sterile seeds and plants were manipulated in a laminar flow cabinet (Fell Clean Air Ltd) to minimize the risk of contamination. Origin of bacteria and preparation of inocula. The three bacterial species, M2, R2 and B 1, were isolated from the rhizospheres of maize, rape (Brassica napus) and barley, respectively. They were identified at the Torry Research Station (Aberdeen, Scotland) as: M2, a fluorescent Pseudomonas sp.; R2, a Mycoplana sp.; B1, a Curtobacterium sp. Our laboratory tests showed that the Pseudomonas sp. (M2) had no growth factor requirements, the Mycoplana sp. (R2) had a requirement for thiamin and the Curtobacterium sp. (Bl) had a requirement for at least one (as yet unidentified) vitamin and amino acid. Inocula were prepared by suspending bacteria from nutrient agar plates (Oxoid) in sterile distilled water to give populations of about 1 x 10" viable cells ml-i. The suspensions were subsequently diluted 1/1000 in 100 ml quantities of sterile PCS ready for addition to the tubes. Foil cap I - I I' I 111 ' Silicone rubber bung ----If# Dl -Serum cap Fig. 1. Apparatus for growing gnotobiotic plants.

3 Bacterial growth in the rhizosphere 97 Inoculation of plants. Seven days after planting, the plants were transferred to a laminar flow cabinet ready for inoculation. Immediately before inoculation the tubes were flushed through with 150 ml sterile distilled water, and the plants were subsequently inoculated by flooding the tubes (from the bottom) with the PCS containing the bacteria. PCS suspension was added until it reached the top of the sand (about 50-60ml). The remaining ml was added from the top of the tube and the sand was allowed to drain freely. This resulted in a dose of about 5 x x lo4 viable bacteria (mg dry wt root)-'. Tubes containing sand but no plants were inoculated in a similar way to serve as controls. Rhizosphere samples and bacterial counts. The number of viable bacteria in the inoculum was estimated by preparing serial 10-fold dilutions in sterile distilled water and making surface colony counts on nutrient agar plates. To estimate the number of organisms in the rhizosphere, roots and closely adhering sand were carefully removed from the tubes. The roots were then divided into three portions: 0 to 2 cm from the tip, 2 to 4 cm from the tip, and 4 cm from the tip to the seed. Each portion was placed in a universal bottle containing 10 ml sterile distilled water and shaken for 10 min on a wrist-action flask shaker (speed setting 7; Stuart Scientific Co., Croydon, Surrey). Serial 10-fold dilutions were subsequently prepared and surface colony counts were made on nutrient agar plates. Counts on control tubes were made by taking samples of about 10 g sand and treating them in the same way as the root portions. Microscopy. Root samples were examined microscopically before and after the washing procedure to determine (a) the pattern and intensity of microbial colonization and (b) whether the washing procedure was effective in removing the bacteria from the root surface. Roots were either mounted directly in 0.5 % (w/v) aqueous phenol/glacial acetic acid (15 :4, by vol.) and, using a Zeiss Universal research microscope (Carl Zeiss, London), examined by Nomaski differential interference contrast, or stained with phenol/aniline blue and examined under bright-field illumination. In the latter method, roots were stained by immersion for 1 min in 0-5 % aqueous phenol/6 % (w/v) aqueous aniline blue/glacial acetic acid (15 : 1 : 4, by vol.), and then carefully rinsed and mounted in 0.5 % aqueous phenol/glacial acetic acid (15 : 4, by vol.). RESULTS Growth rates of single species The maximum growth rates and final populations achieved by the individual species in the different rhizospheres were calculated on the basis of the number of viable bacteria per mg dry wt root (Table 1). The pattern of development of the Mycoplana sp. (R2) in the rhizosphere of wheat (Fig. 2) typified that of all three bacterial species (Pseudomonas, Mycoplana, Curtobacterium) in the rhizospheres of wheat and barley. The populations in the middle and older parts of the root system were always similar but in the tip region, the population was consistently smaller by a factor of about 10. This was probably because new tissue was continually being added to this region during the course of the experiment and so there was no time for the bacterial population to become as dense as in the areas of more mature tissue. The development of the Curtobacterium sp. and the Mycoplana sp. in the rhizosphere of maize was different from that in the rhizosphere of wheat or barley (Fig. 3). There was an initial lag period of about 28 h before the number of either organism increased. The number of Mycoplana sp. increased more slowly and reached a smaller final population than when in the rhizosphere of wheat or barley. The Curtobacterium sp. increased slightly in numbers from 28 to 48 h after inoculation but subsequently declined slowly, suggesting that this species was not suited for growth in the maize rhizosphere. In the control tubes (those without plant roots) the numbers of all three species increased two- to fourfold during the first h after inoculation but then declined slowly. Growth rates of mixed populations When the Curtobacterium sp. and the Mycoplana sp. were co-inoculated into the rhizospheres of wheat or barley, the growth rates and final populations of both strains were slightly less than when grown singly (Table 1). When the Mycoplana sp. was co-inoculated with the Curtobacterium sp. into the rhizosphere of maize, it developed in the same way as when it was inoculated alone (Fig. 3). However, the growth of the Curtobacterium sp. in the

4 98 R. A. BENNETT AND J. M. LYNCH Table 1. Maximum growth rates and final populations of three bacterial species alone and co-inoculated in the rhizosphere of wheat, barley and maize Curtobacterium sp. (B I), Mycoplanu sp. (R2) and Pseudomonas sp. (M2) were inoculated separately or together, as indicated. For each --- bacterial species, the doubling time, td (h), during the exponential growth phase of the population and the final population, x, [no. of viable bacteria (mg dry wt root)-'], were determined. The results shown are means for the parts of the root 2 to 4 cm from the tip and 4 cm from the tip to the seed base. Wheat Barley Maize Bacterial species 'd Xmax 'd Xmax td X,*X B1 alone x lo x lo6 NG NG R2 alone lo6 ~ x lo x 105 M2 alone lo6 ~ x lo6 NT NT B 1 (co-inoculated with R2) lo6 ~ x lo x 105 R2 (co-inoculated with B 1) x lo x lo6 5.2 OX 105 B 1 (co-inoculated with R2, M2) NG NG NT NT NT NT R2 (co-inoculated with B 1, M2) ~ NT NT NT NT M2 (co-inoculated with B 1, R2) ~ NT NT NT NT lo6 NG, No growth; NT, not tested T Time after inoculation (d) Fig. 2. Growth of a Mycoplanu sp. in different parts of the wheat rhizosphere: 0 to 2 cm from root tip (0); 2 to 4 cm from tip (0); 4 cm from tip to seed base (Q Time after inoculation (d) Fig. 3. Growth of a Curtobacterium sp. and a Mycoplanu sp. in the maize rhizosphere when inoculated separately and together: Curtobacterium sp. alone (0); Mycoplanu sp. alone (@); Curtobacterium sp. (0) and Mycoplana sp. (.) co-inoculated.

5 Bacterial growth in the rhizosphere 99 \ I 1 I I I Time after inoculation (d) Fig. 4. Growth of a Curtobacterium sp. (O), a Pseudomonas sp. (0) co-inoculated in the wheat rhizosphere. and a Mycoplanu sp. (0) when maize rhizosphere was stimulated by the presence of the Mycoplana sp., its final population reaching 10 times that when it was inoculated alone. When all three species were co-inoculated into the rhizosphere of wheat, we observed large differences in the growth of each species compared with when they were inoculated alone (Table 1, Fig. 4). The growth rate of the Pseudomonas sp. during the early stages was stimulated, though the final population was similar to when it was inoculated alone (Table 1). In contrast, the growth rate of the Mycoplana sp. was retarded and the final population was only one-tenth of that when inoculated alone. The most marked effect was on the Curtobacterium sp. which increased slightly in numbers during the first 24 h after inoculation but then disappeared completely: thus it was a poor competitor under the conditions of these experiments. Microscopy No bacteria were observed on the root surface after shaking, suggesting that the procedure used was effective in removing these particular species. The Pseudomonas sp. was abundant on the surface of wheat roots 7 d after inoculation [about 7 x lo7 bacteria (mg dry wt root)-'] (Figs 5-8); similar observations were made when this organism was inoculated into the barley rhizosphere. Even careful washing of the roots in water to remove sand particles also removed many of the bacteria. Large sand particles were therefore removed with forceps and then the root Was immersed directly into the stain. Bacteria were found to aggregate along the length of the root apparently in the intercellular spaces (Fig. 5), or in discrete microcolonies some of which appeared to be on the surface of root cells and not in the intercellular spaces (Fig. 6). The appearance of many single cells dispersed over the whole root surface (Fig. 7) was possibly caused by the preparation procedure disrupting the aggregates and colonies of bacteria, since we did sometimes observe partially disrupted microcolonies (Fig. 8). Similar observations were made using Nomaski differential interference contrast; for example, lines of bacteria in the intercellular spaces of the epidermis (Fig. 9) and the formation of a colony across the surface of epidermal cells (Fig. 10). We did not measure what fraction of the root surface was covered but it was small and varied greatly, even between areas only a few micrometres apart. In similar preparations of roots inoculated with the Curtobacterium sp. and the Mycoplana sp. no colonies were seen attached to the root surface though many bacteria were observed floating in the surrounding solution.

6 100 R. A. BENNETT AND J. M. LYNCH

7 Bacterial growth in the rhizosphere 10 1 Effect of bacteria on plant growth The Curtobacterium sp. and the Mycoplana sp. had no obvious effect on the growth of wheat, barley or maize. However, wheat and barley plants inoculated with the Pseudomonas sp. began to become chlorotic 5-6 d after inoculation, starting at the base of the first two leaves and progressing successively upwards. DISCUSSION That the Curtobacterium sp. was able to grow in the rhizosphere of wheat and barley, but not in the rhizosphere of maize, indicates specificity between bacteria and plant species. Two possible explanations for the lack of growth in the maize rhizosphere are that the substrates available around maize roots lack a particular growth factor (e.g. vitamin, amino acid) required by the Curtobacterium sp. or that the compounds released by the roots are inhibitory or toxic to the bacteria. That the organism showed a slight increase in numbers and did not die out completely suggests nutrient deprivation rather than toxins. Bowen (1979) obtained similar results, including evidence for specificity, using three strains of Pseudomonas fluorescens. Two of the strains had quite dissimilar growth rates and one was unable to grow in the rhizosphere of Eucalyptus seiberi, for reasons that remain unclear. The slight decrease in growth rate and final population of the Curtobacterium sp. and the Mycoplana sp. when co-inoculated into the rhizosphere of wheat or barley suggests that both organisms were competing for the same nutrients or possibly for the same sites in the rhizosphere. The cumulative growth rate and final population of the two strains were the same as when either strain was inoculated alone, suggesting that the amount of microbial growth was determined by the quantity of substrates released by the roots. In contrast, when the Curtobacterium sp. was co-inoculated with the Mycoplana sp. into the rhizosphere of maize, the Curtobacterium sp. was stimulated and increased significantly in population compared with when it was inoculated alone. The growth and development of the Mycoplana sp. appeared unaffected (commensalism). Perhaps the Mycoplana sp. stimulated the maize root to release different compounds, or the metabolites of the Mycoplana sp. may contain the growth factors required by the Curtobacterium sp. The growth curves for all three bacterial species used here in wheat and barley were similar to those reported by Bowen & Rovira (1973) for the total soil bacteria and pseudomonads in the rhizosphere of Pinus radiata. When all three bacterial species were co-inoculated into the wheat rhizosphere two degrees of antagonism were evident. One possible explanation is offered by the microscopic evidence. The Pseudomonas sp. seemed to be more intimately associated with the root than either of the Figs Root surface of wheat inoculated with a Pseudomonas sp. The photographs were taken 7 d after inoculation when the bacterial population (estimated by surface colony counts) was about 7 x lo7 (mg dry wt root)-'. Roots were either stained with phenollaniline blue (Figs 5-8) or mounted unstained in phenol/acetic acid and observed by Nomaski differential interference microscopy (Figs 9, 10). All the bar markers represent 5 pm. Fig. 5. Aggregates of bacteria in the intercellular spaces. Fig. 6. Discrete microcolonies distributed across the surface of epidermal cells. Fig. 7. Mass of single cells apparently distributed over the root surface. (This effect was thought to be caused by the preparation procedure.) Fig. 8. Partially disrupted microcolony. Fig. 9. Aggregates of bacteria in the intercellular spaces. Fig. 10. Microcolony spreading across the surface of epidermal cells.

8 102 R. A. BENNETT AND J. M. LYNCH other species in that it alone developed microcolonies on the rhizoplane, presumably affording it the advantage of being closest to the substrate supply. Although there was no evidence of antibiotic production by the Pseudomonas sp. or the Mycoplana sp. on nutrient agar plates, the possibility of antibiosis cannot be rejected without qualitative and quantitative details of the root exudate composition. The observation that there was little effect on the Curtobacterium sp. when it was co-inoculated with the Mycoplana sp. suggests that it was the presence of the pseudomonad which so adversely affected the growth of the Curtobacterium SP. The preferential colonization of cell junctions by microbes was first observed by Rovira (1956) and subsequently verified by experiment (Bowen, 1979). However, our results show microcolony development on the surface of epidermal cells as well as at the cell junctions. Are the cell junctions really the major sites of release of carbon by roots as suggested by Bowen (1979) or are they sites of accumulation? Do more colonies develop in the cell junctions or are the colonies here less easily disrupted during slide preparation? We believe that further studies using the defined systems will lead to a better understanding of the complex environment of the rhizosphere. We thank Miss Susan Medhurst for excellent technical assistance. REFERENCES ASANUMA, S., TANAKA, H. & YATAZAWA, M. (1979). Rhizoplane micro-organisms of rice seedlings as examined by scanning electron microscopy. Soil Science and Plant Nutrition 25, BARBER, D. A. (1967). The effect of micro-organisms on the absorbtion of inorganic nutrients by intact plants. I. Apparatus and culture technique. Journal of Experimental Botany 18, BARBER, D. A. & LYNCH, J. M. (1977). Microbial growth in the rhizosphere. Soil Biology and Biochemistry 9, BERKELEY, R. C. W., LYNCH, J. M., MELLING, J., RUTTER, P. R. & VINCENT, B. (editors) (1980). Microbial Adhesion to Sursaces. Chichester: Ellis Horwood. BOWEN, G. D. (1979). Integrated and experimental approaches to the study of growth of organisms around roots. In Soil-borne Plant Pathogens, pp Edited by B. Schippers & W. Gams. London: Academic Press. BOWEN, G. D. (1980). Misconceptions, concepts and approaches in rhizosphere biology. In Contemporary Microbial Ecology, pp Edited by:d. C.. Ellwood, J. N. Hedger, M. J. Latham, J. M. Lynch & J. H. Slater. London: Academic Press. BOWEN, G. D. & ROVIRA, A. D. (1973). Are modelling approaches useful in rhizosphere biology? Bulletin of the Ecological Research Commission, Natural Science Research Council, Sweden 17, DART, P. J. (197 1). Scanning electron microscopy of plant roots. Journal of Experimental Botany 22, ELLIOTT, L. F., GILMOUR, C. M., LYNCH, J. M. & TITTEMORE, D. (1981). Bacterial colonisation of plant roots. In Microbial-Plant Interactions, (in the Press). Edited by R. L. Todd. Madison: Soil Science Society of America. LYNCH, J. M. (1981). The rhizosphere. In Experimental Microbial Ecology, (in the Press). Edited by R. G. Burns & J. H. Slater. Oxford: Blackwell Scientific Publications. NEWMAN, E. I. (1978). Root micro-organisms: their significance in the ecosystem. Biological Reviews 53, NEWMAN, E. I. & WATSON, A. (1977). Microbial abundance in the rhizosphere. A computer model. Plant and Soil 48, ROVIRA, A. D. (1956). A study of the development of the root surface microflora during the initial stages of plant growth. Journal of Applied Bacteriology 19, ROVIRA, A. D., NEWMAN, E. I., BOWEN, H. J. & CAMPBELL R. (1974). Quantitative assessment of the rhizoplane microflora by direct microscopy. Soil Biology and Biochemistry 6,