Biol Fertil Soils (2000) 30:450-459

Y. Bashan · P. Vazquez

Effect of calcium carbonate, sand, and organic matter levels on mortality of five species of Azospirillum in natural and artificial bulk soils Received: 9 April 1999

Abstract Five bacterial strains, one from each of the five known species of the plant growth-promoting bacteria (PGPB) Azospirillum (A. brasilense, A. lipoferum, A. amazonense, A. halopraeference, and A. irakense) were inoculated into two natural, semiarid soils (terra rosa and loessial sandy) from Israel, and two artificial soils constructed to simulate the native soils. Within 60 days, the populations of all five Azospirillum species declined significantly in a linear fashion, in both the native soils and in the homologous artificial soils. Increased levels of CaCO3, and fine and rough sand, had significant detrimental effects on the survival of the five Azospirillum species, whereas increased organic matter content improved survival. In contrast, when the bacterial strains were incubated in the rhizosphere of tomato seedlings grown in the artificial soils, manipulation of these soil variables had only a marginal effect on bacterial survival; all Azospirillum species survived well in the tomato rhizosphere under conditions that are otherwise detrimental. This study indicates that most cells of the strains of five known species of Azospirillum died out linearly over time in two semiarid soils, and that only the major soil components affected Azospirillum survival in soil. Because mortality was similar in native soils and in artificial homologous soils, artificial soils can be used to study the soil behavior of Azospirillum. Dedicated to Prof. K. Vlassak on the occasion of his 65th birthday and to the memory of the late Mr. Avner Bashan of Israel Y. Bashan ()) ⋅ P. Vazquez Environmental Microbiology, The Center for Biological research of the Northwest (CIB), La Paz, PO. Box 128, Baja California Sur 23000, Mexico e-mail: bashan@cibnor.mx Tel.: + 112-536-33 ext. 200 Fax: + 112-547 10

Key words Azospirillum · Bacterial mortality · Plant-growth-promoting bacteria · Plant-growth-promoting rhizobacteria · Bacterial survival Introduction Azospirillum species are important rhizosphere bacteria. They proliferate in the rhizosphere of numerous plant species of different families and have been shown to stimulate the growth of associated plants (Bashan and Holguin 1997), as do many other plant-growth-promoting bacteria (PGPB)s (Bashan and Holguin 1998; Glick 1995). However, little is known about the behavior of Azospirillum in bulk soil, e.g., whether it is able to multiply in the absence of a host plant (Del Gallo and Fendrik 1994).

Azospirillum has been isolated from the roots and rhizospheres of numerous host plants [for a review see Bashan and Holguin (1997)], and has been inoculated into especially agriculturally important annuals such as cereals (Jagnow 1987) and vegetables (Bashan et al. 1989). The practice of harvesting these crops, which are the source of nutrients for bacteria, from the soil at the end of each growing season, creates a hurdle for the survival of Azospirillum from one season to the next. To re-establish a population in the following season, the bacteria must be able to survive, and perhaps proliferate, in soils in the absence of the host plant.

Although Azospirillum species have been successfully isolated from bulk soil (Bashan 1999), from the beginning of agricultural research on these species, the capacity of these bacteria to survive in bulk soil has been shown to be contradictory. In some soils, they survived well for long periods (Baldani et al. 1986; Bouillant et al. 1997; Kabir et al. 1994; Oliveira and Drozdowicz 1988); however, in others they did not (Albrecht et al. 1983; Bashan and Levanony 1987, 1988; Bashan et al. 1991a), although DNA of

© Springer-Verlag 2000

451

Azospirillum species (16s sequences) is always present in soils (R. Bally, personal communication).

The effects of single soil variables, such as soil texture and clay content, on bacterial survival are known (van Elsas 1992; Hartel et. al. 1994). In Israeli and Mexican soils, the content of clay, N, and organic matter, and the water-holding capacity were positively correlated with A. brasilense viability. A high percentage of CaCO3, and fine or rough sand, had a negative effect on its viability in soil (Bashan et al. 1995).

The aims of this study were: (1) to extend the results of survival in bulk soil obtained earlier with A. brasilense to the other four species of Azospirillum, and (2) to determine whether members of the genus Azospirillum in general are affected by the levels of CaCO3, sand, and organic matter in soil. In addition to using natural soils, artificially constructed soils were evaluated for their ability to exert an analogous effect on bacterial survival and, therefore, for their use in studying Azospirillum behavior in bulk soil. Materials and methods Organisms and growth conditions The following Azospirillum strains were used: A. brasilense Cd (DSM 7030, Braunschweig, Germany), A. lipoferum 1842 (DSM 1842), A. halopraeference Au4 (donated by B. Reinhold-Hurek, Max Plank Institute, Marburg, Germany), A. amazonense 2787 (DSM 2787), and A. irakense KBC1 (donated by P. Kaiser, Institute National Agro nomique, Paris, France). Tomato (Licopersicon esculentum Mill. cv. VF M-82-1-8) seedlings were used as host plants.

Azospirillum strains were grown on N-free OAB medium (Bashan et al. 1993) for 16 h at 30±1 °C at 200 rpm. The medium for growing A. halopraeference Au4 was as described by Reinhold et al. (1987). All bacteria were harvested by centrifugation (7000 g, 10 min), washed twice in K3P04 buffer (PBS, pH 7.0, 0.06 M) supplemented with 0.15 M NaCl, and repared for soil inoculation at a final concentration of 1 x 108 colony forming units (cfu)/ml soil (Bashan 1986). Soil was directly inoculated by adding a triple -washed bacterial suspension in sterile water to each pot. Plants were inoculated by a standard seed treatment (Puente and Bashan 1993) and again immediately after germination (after emergence of cotyledons from the soil surface) by soil inoculation as described above. Soils Two native soils from Israel evaluated in a previous study (Bashan et al. 1995) were used: terra rosa soil (Rhodoxeralfs) and loessial sandy soil (Torripsamments). Their characteristics are given in Table 1. Artificial soil

Soil is an extremely difficult environment to simulate. Nevertheless, to evaluate whether changes in relative quantities of soil components affect the s urvival of Azospirillum species, two artificial soils were created. They were prepared to resemble, in their major components, the natural soils from Israel mentioned above. Both artificial soils were composed of quartz sand (fine, particle size 0.02-0.2 mm; rough, particle size 0.2-2 mm) thoroughly washed with distilled water and other components in the proportions described for the natural soil in Table 1. The two primarily clay minerals in each soil were added in 1:2 (v/v) proportions to produce the required clay concentrations (montmorillonite and kaolinite for Rhodoxeralfs and montmorillonite and calcite for Torripsamments). Ca was added as analytical-grade CaCO3, and organic matter was added as milled (60 mesh), thoroughlywashed sawdust. The water-hold ing capacity was adjusted by the addition of very fine vermiculite (particle size < 1 mm). Silt with an average particle size of 0.02 mm was donated by the local aquaculture industry. The pH of the soil was adjusted with phosphate buffer, to pH 7.0-7.8 (which served also as a supply of K and P), and the N content was adjusted by adding NH4NO3. No microelements were added as they were present as contaminants in the analytical-grade reagents used. All the ingredients were mixed in a small, homemade soil mixer. Common soil analyses of the resultant artificial soils revealed that their basic physical and chemical characteristics were similar to those of the original soils upon which they were based. Soil moisture retention curves were compared and were also similar.

Sand concentrations were manipulated by adjusting the relative percentages of fine and rough sand while keeping the total level of sand in the artificial soil at the same level as in the native soil. Soil mixtures, sterilized by the standard autoclave tyndelization method (Bashan et al. 1995), were aliquoted (100 ml) into black plastic pots and irrigated to water-holding capacity with deionized water. Plant growth conditions and soil incubation Five tomato seeds were sown in natural and artificial soils in pots (150 ml containing 100 ml soil) in a Conviron TG-16 growth chamber (28 ± 2 °C, 12 h light with a total light intensity of 200 µmol/m2/s; Conviron, Winnipeg, Canada) for 3 weeks. Plants were grown either in natural soils, in homologous artific ial soils, or in two artificial soils in which CaCO3 and sand levels were modified. In the latter soils, the level of CaC03 was increased to 3% (v/v), the level of rough sand was adjusted to 9% (v/v), and the level of fine sand to 90% (v/v), eliminating the clay and silt fractions (modified artificial loessial sandy soil). In modified artificial terra rosa soil the level of CaCO3 was increased to 3% (v/v), the level of rough sand was adjusted to 5% (v/v), and the level of fine sand to 50% (v/v), and the level of clay was reduced to 18% (v/v). To avoid nutritional stress to the plants (indicated by chlorosis), a commercial complete fertilizer for gardens (0.5% w/v) was added twice by irrigation after 3 and 5 weeks. Plants were irrigated with 10 ml distilled water 3 times a week. Bulk soil (containing no plants) was incubated and irrigated similarly.

452 Quantification of bacterial in bulk and rhizosphere soil Soil samples for bacterial quantification from bulk and rhizosphere soils were prepared as previously described (Bashan et al. 1995). The level of bacteria in the soil samples was determined by a basic enzyme-linked immunosorbent assay (ELISA) technique (Levanony et al. 1987) when the level of bacteria was higher than 105 cfu/sample, and by the limited enrichment technique coupled with ELISA detection (Bashan et al. 1991b) when fewer bacteria were present. Results are presented as cfu/g soil.

Experimental design and statistical analysis Soil samples for determination of bacterial survival in both natural and artificial soils were taken randomly with five replicates for each; each pot was considered a re plicate. To determine the level of bacteria in both natural and artificial soils, five 0.5-g soil samples were taken randomly from a single pot, at each sampling time, and thoroughly mixed, creating one homogeneous sample of 2.5 g per sampling time. From the homogeneous sample, nine subsamples (0.1 g each) were taken randomly for bacterial quantification, yielding initially nine values per pot. Values for three subsamples

were combined to finally yield three values per pot; these final values were subjected to statistical analysis. This process was repeated for each of five pots of identical soil composition and bacterial treatment in a single experiment (n =100 pots for all Azospirillum species). The experiment with artificial soils was repeated twice and results presented are the averages of these. Linear regressions were made by SigmaPlot 5.0 software (Jandel, San Rafael, Calif.) with significance calculated by the Spreadware statistics package (Spreadware, Palm Desert, Calif.). Actual P values are given in the figures.

Results Survival of Azospirillum in two natural and two artificial bulk soils In both natural soils, the populations of the five species of Azospirillum declined linearly (P ≤ 0.001) and rapidly after inoculation, with only small differences between the species (Fig. 1A-J, solid lines). In the loessial

453

454

sandy soil, populations of the five Azospirillum species declined even more rapidly, and linearly (P ≤ 0.001) than in the terra rosa soil. Survival among the species in both natural soils was as follows: A. irakense>A. halopraeference>A. brasilense>A. amazonense>A. lipoferum. The mortality rates of the five Azospirillum species in the two artificial soils were almost identical to those in the homologous natural soils (Fig. 1A-J, dashed lines).

Combining all of the data from the five Azospirillum species for survival over time in each soil type (natural or artificial) revealed that populations of each species of Azospirillum declined linearly, similarly, and significantly (P ≤ 0.001) in both natural and artificial homologous soils (Fig. 2A-D).

The effect of addition of CaCO3, sand, and organic matter to artificial soils on survival of Azospirillum The three Azospirillum species with the highest survival, A. irakense, A. halopraeference, and A. brasilense, were further evaluated for survival in two artificial soils with elevated CaCO3 concentrations (1-5%). The original soils had a CaCO3 concentration of 0.5% (terra rosa) and 1.6% (loessial sandy) (Table 1).

Increased levels of CaCO3 in both artificial soils significantly and linearly decreased survival of all three species after 14 days (Fig. 3). Very few bacteria survived when the level of CaCO3 reached 5%; only single cells up to approximately 101 cfu/g soil were detected in these soil samples.

The level of fine sand in terra rosa soil is 27.4% and in the loessial sandy soil 88.4% (Table 1). In these soils, the population of the three Azospirillum species ranged from 103-105 cfu/g after 14 days. Increasing the per

455

centage of fine sand in the artificial terra rosa soil significantly and linearly reduced survival in all three species (Fig. 4a,c,e). In the artificial loessial sandy soil, reducing the original high percentage of fine sand increased survival (Fig. 4b,d,f). Increasing the level of fine sand in the artificial soil from 88% to 100% only slightly reduced survival.

The level of rough sand in natural terra rosa soil is 1.6% and in the loessial sandy soil 8.3% (Table 1). At these concentrations, 103-104 cfu/g soil survived after 14 days. Increasing the percentage of rough sand in both artificial soils up to 20% significantly and linearly reduced survival of all three species (Fig. 5).

The level of organic matter in both natural soils was low; terra rosa soil contained 2.8% and in the loessial sandy soil it was only 0.3% (Table 1). Increasing the percentage of organic matter in both artificial soils up to 5%, significantly and linearly increased survival of all the

Azospirillum species, reached 108 cfu/g after 14 days in soil containing 5% organic matter (Fig. 6) Survival of Azospirillum in tomato rhizosphere with high levels of sand and CaCO3 Tomato seedlings grew well in both artificial soils containing high levels of CaCO3 and sand; seedlings developed to the one- or two-true-leaf stage. At this stage, the seedling roots supported significant numbers of all Azospirillum species similar to the levels of bacteria supported by seedlings grown in natural soils, or homologous artificial soils without additional CaCO3 or sand (Table 2). Beyond this stage of development, plants could only grow when complete fertilizer was added, though the rhizosphere population of Azospirillum species did not change (Table 2).

456

Discussion Plants are usually inoculated with PGPB either by seed treatment, or more commonly, by application of bacteria to the soil (as a liquid suspension or incorporated with an inoculant carrier) as near as possible to the germinating seedlings. The site of inoculation likely influences the fate of a bacterium, i.e., whether the bacterium colonizes the plant roots or dies out. Persistence of Azospirillum in the soil is crucial for plant growth promotion because seed inoculation is impractical in many situations, e.g., for establishing a population of bacteria on perennial plants from one season to the next, for inoculating vegetatively propagated plants and trees, and when more than one inoculation per season is required (Bashan 1998). In bulk soil, Azospirillum cells are subject to natural physicochemical forces, to interactions

with soil particles including adsorption, and encapsulation by clay minerals, and to the wet and dry regimes of the soil.

Although Azospirillum has been isolated from many soils (Bashan 1999), the presence of the bacteria in the soil is not necessarily evidence that the cells are physiologically active in situ. On the contrary, nutrient limitation is a major stress factor in soil and thus these bacteria may endure long periods of inactivity and sparse periods of growth (Stotzky 1997).

Although the physiological state of Azospirillum cells may affect their survival in the soil (Vandenhove et al. 1993), and A. lipoferum strain CRT1 can survive in the soil independently of the soil type (Jacoud et al. 1998), in general, a change in the physicochemical properties of the soil may have a much greater effect on bacterial survival in the soil than changes of the bacteria itself (Stotzky 1997). In Brazilian and Canadian

457

soils, the population of A. brasilense Cd declined rapidly within the first 2 weeks after inoculation, and finally established a constant level of about 105 cfu/g soil. Nutrient amendment caused the bacterial population to increase for several days, after which it returned to its initial level (Germida 1986). The survival of A. brasilense Cd and A. brasilense Sp-245 in 23 types of plantfree sterilized soils, obtained from a wide range of environments in Israel and Mexico, showed that survival was mainly related to the soil type and not whether the soil came from an arid zone. In several soils from Israel, the populations of two A. brasilense strains rapidly decreased with time (Bashan et al. 1995). The study presented here extends this earlier research to include other species of the genus Azospirillum; only small differences in mortality rates existed between the species in a particular soil. Thus,

survival is related to the soil type and not to the bacterial species.

Viability of A. brasilense in two artificial soils containing the same major components as the natural soils from Israel was almost identical to that in the natural soils used (Bashan et al. 1995). The present study on the scope of soil survival of the genus Azospirillum showed that other species of Azospirillum behave similarly in natural and artificial soils. Furthermore, this study showed that only the major soil variables influence Azospirillum survival, and not the minor components such as the soil pore space. This is similar to the behavior of Rhizobium leguminosarum bv. trifolii in soil (Postma and van Veen 1990). The earlier findings that elevated levels of CaCO3, and fine and rough sand are detrimental to A. brasilense survival, and that addition of organic matter enhances A. brasilense survival (Bashan and Levanony 1988; Bashan et al. 1995) can be extended to all species of Azospirillum. Finally, this study showed that artificial soil environments can be

457

458

constructed and controlled relatively easily, and that, at least in the case of Azospirillum, artificial soils are suitable for studying interactions between soil and PGPBs.

The results presented here might have important practical implications. Because of the enormous variability of soils even on a small scale (van Elsas 1992), conventional wisdom recommends checking each strain destined for field trials for their ability to survive in relevant soils. Sometimes field trials involve numerous strains and evaluation of soil survival can be very laborious and time consuming. On the basis of our

findings, it may not be necessary to check the soil survival behavior of each strain; one or two representative strains may be sufficient.

In summary, this study showed that, in contrast to survival in the rhizosphere, most cells of the five known species of Azospirillum died out within 60 days in two semiarid bulk soils. Mortality was linear and similar in native soils and in artificial homologous soils, and thus, artificial soils can be used as models for studying the soil behavior of Azospirillum. Only the major soil components

459 affected Azospirillum survival; CaCO3 and sand had a detrimental effect on survival and organic matter enhanced survival. Acknowledgements This study was supported by Consejo Nacional de Ciencia y Tecnologia (CONACyT), Mexico, contracts no. 26262-B and no. 28362-B. We thank Mrs. Cheryl Patten for critically reading and styling the manuscript and Dr. Ellis Glazier for editing the English text. References Albrecht SL, Gaskins MH, Milam JR, Schank SC, Smith RL (1983)

Ecological factors affecting survival and activity of Azospirillum in the rhizosphere. Exper Suppl 48:138-148

Baldani VLD, Alvarez MA de B, Baldani JI, Döbereiner J (1986) Establishment of inoculated Azospirillum spp. in the rhizosphere and in roots of field grown wheat and sorghum. Plant Soil 90:35-46

Bashan Y (1986) Significance of timing and level of inoculation with rhizosphere bacteria on wheat plants. Soil Biol Biochem 18:297-301

Bashan Y (1998) Inoculants of plant growth-promoting bacteria for use in agriculture. Biotechnol Adv 16:729-770

Bashan Y (1999) Azospirillum spp. interactions in soils - a review. Biol Fertil Soils 29:246-256

Bashan Y, Holguin G (1997) Azosp irillum-plant relationships: environmental and physiological advances (1990-1996). Can J Microbiol 43 :103-121

Bashan Y, Holguin G (1998) Proposal for the division of plant growth-promoting rhizobacteria into two classifications: biocontrol-PGPB (plant growth-promoting bacteria) and PGPB. Soil Biol Biochem 30:1225-1228

Bashan Y, Levanony H (1987) Horizontal and vertical movement of Azospirillum brasilense Cd in the soil and along the rhizosphere of wheat and weeds in controlled and field environments. J Gen Microbiol 133:3473-3480

Bashan Y, Levanony H (1988) Adsorption of the rhizosphere bacterium Azospirillum brasilense Cd to soil, sand and peat particles. J Gen Microbiol 134:1811-1820

Bashan Y, Ream Y, Levanony H, Sade A (1989) Nonspecific responses in pla nt growth, yield, and root colonization of noncereal crop plants to inoculation with Azospirillum brasilense Cd. Can J Bot 67:1317-1324

Bashan Y, Mitiku G, Whitmoyer RE, Levanony H (1991a) Evidence that fibrillar anchoring is essential for Azospirillum brasilense Cd attachment to sand. Plant Soil 132:73-83

Bashan Y, Mitiku G, Ziv-Vecht O, Levanony H (19916) Estimation of minimal numbers of Azospirillum brasilense using time-limited liquid enrichment combined with enzyme-linked immunosorbent assay. Soil Biol Biochem 23:135-138

Bashan Y, Holguin G, Lifshitz R (1993) Isolation and characterization of plant growth-promoting rhizobacteria. In: Glick BR, Thompson JE (eds) Methods in plant molecular biology and biotechnology. CRC Press, Boca Raton, Fla., pp 331-345

Bashan Y, Puente ME, Rodriguez-Mendoza MN, Toledo G, Holguin G, Ferrera-Cerrato R, Pedrin S (1995) Survival of Azospirillum brasilense in the bulk soil and rhizosphere of 23 soil types. Appl Environ Microbiol 61:1938-1945

Bouillant ML,Miche L, Ouedraogo O, Alexandre G, Jacoud C, Salle G, Bally R (1997) Inhibition of Striga seed germination associated with sorghum growth promotion by soil bacteria. CR (Dok) Acad Sci USSR 320:159-162

Del Gallo M, Fendrik I (1994) The rhizosphere and Azospirillum. In: Okon Y (ed) Azospirillum/plant associations. CRC Press, Boca Raton, Fla., pp 57-75

Elsas JD van (1992) Environmental pressure imposed on GEMMOS in soil. In: Stewart -Tull DES, Sussman M (eds) The release of genetically modified microorganisms. Plenum Press, New Yo rk, pp 1-14

Germida JJ (1986) Population dynamics of Azospirillum brasilense and its bacteriophage in soil. Plant Soil 90:117-128

Glick BR (1995) The enhancement of plant growth by free-living bacteria. Can J Microbiol 41:109-117

Hartel PG, Fuhrmann JJ, Johnson WIT Jr, Lawrence EG, Lopez CS, Mullen MD, Skipper HD, Staley TE, Wolf DC, Wollum AG II, Zuberer DA (1994) Survival of a lacZY-containing Pseudomonas putida strain under stressful abiotic soil conditions. Soil Sci Soc Am J 58:770-776

Jacoud C, Faure D, Wadoux P, Bally R (1998) Development of a strain -specific probe to follow inoculated Azospirillum lipoferum CRTl under field conditions and enhancement of maize root development by inoculation. FEMS Microbiol Ecol 27:43-51

Jagnow G (1987) Inoculation of cereal crops and forage grasses with nitrogen-fixing rhizosphere bacteria: possible causes of success and failure with regard to yield response - a review. Z Pflanzenernaehr Bodenkd 150:361-368

Kabir M, Chotte JL, Rahman M, Bally R, Jocteur Monrozier L (1994) Distribution of soil fractions and location of soil bacteria in a vertisol under cultivation and perennial grass. Plant Soi1163:243-255

Levanony H, Bashan Y, Kahana ZE (1987) Enzyme-linked immunosorbent assay for specific identification and enumeration of Azospirillum brasilense Cd in cereal roots. Appl Environ Microbiol 53 : 358-364

Oliveira RGB, Drozdowicz A (1988) Are Azospirillum bacteriocins produced and active in soil? In: Klingmüller W (ed) Azospirillum IV: genetics, physiology, ecology. Springer, Berlin Heidelberg New York, pp 101-108

Postma J, Veen JA van (1990) Habitable pore space and survival of Rhizobium leguminosarum biovar trifolii introduced into soil. Microb Ecol 19:149-161

Puente M-E, Bashan Y (1993) Effect of inoculation with Azospirillum brasilense strains on the germination and seedlings growth of the giant columnar cardon cactus (Pachycereus pringlei). Symbiosis 15:49-60

Reinhold B, Hurek T, Fendrik I, Pot B, Gillis M, Kersters K, Thielemans S, De Ley J (1987) Azospirillum halopraeference sp. nov., a nitrogen-fixing organism associated with roots of Kallar grass (Leptochloa fusca (L.) Knuth). Int J Syst Bacteriol 37 : 43-51

Stotzky G (1997) Soil as an environment for microbial life. In: Elsas JD van, Trevors JT, Wellington EMH (eds) Modern soil microbiology. Dekker, New York, pp 1-20

Vandenhove H, Merckx R,Steenbergen M van, Vlassak K (1993) Microcalorimetric characterization, physiological stages and survival ability of Azospirillum brasilense. Soil Biol Biochem 25:513-519