RESEARCH WORK

Isolation and identification of Azospirillum sp. in Guinea grass (Panicum maximum Jacq.) of the Valle del Cesar

Diana M. Cárdenas¹, María F. Garrido², Ruth R. Bonilla³ y Vera L. Baldani⁴

E-mail: dicarcaro@hotmail.com

²Corporación Colombiana de Investigación Agropecuaria-CORPOICA. Bogotá, Colombia

³Laboratorio de Microbiología de Suelos Corporación Colombiana de Investigación Agropecuaria-CORPOICA. Bogotá, Colombia

⁴EMBRAPA Agrobiología. Seropedica-Brasil

ABSTRACT

The effect of environmental factors of the Valle del Cesar and the agronomic management of Guinea grass (Panicum maximum Jacq.), on the bacterial population of the Azospirillum genus in semisolid NFb and LGI culture media was evaluated, for which an experimental design was used in divided plots with a 2 (climatic seasons: rainy and dry) x 2 (agronomic managements: agroecological and extractive) x 3 (analyzed samples: rhizospheric soil, roots and leaves) factorial arrangement. The results did not reveal significant statistical differences, which indicates that this bacterium can maintain its population under stress conditions by different physiological mechanisms. From these samples 16 isolations were obtained belonging to the Azospirillum genus in which their acetylene-reduction activity was evaluated as indicator of biological nitrogen fixation and their capacity in the production of indolic compounds as plant growth promoters. The strains SRGM2, SRGM3 and SRGM4, obtained from rhizospheric soil samples of Guinea grass of the Experimental Station Motilonia of Corpoica, Agustín Codazzi municipality, Cesar department, were selected. These isolations were molecularly characterized by the gen 16S rRNA and according to the BLAST analysis in the GenBank database and showed 93% similarity with A. lipoferum (SRGM2 and SRGM3) and 94% with A. brasilense (SRGM4).

Key words: Growth, nitrogen fixation, Azospirillum sp.

INTRODUCTION

Cesar is one of the main livestock production departments of the Colombian Caribbean region and constitutes an important milk-producing region (Gamarra, 2005). Its soils are derived from well-drained sedimentary materials and the pH is close to neutrality; they have low organic matter levels, high Ca contents and low levels of Mg, K and minor elements; its texture varies from loamy to loamy-clayey. Forage species are cultivated on them, among which Guinea grass (Panicum maximum Jacq.) reaches high productivity in dry matter, around 12 to 18 t/ha (Cuesta-Muñoz et al., 2005). However, these soils show physical, chemical and biological deterioration, which severely affects their productive capacity and compromises the economic viability of livestock production systems. The above-expressed is caused, among other factors, by the indiscriminate use of chemical fertilizers, mainly as source of N, which plays an important role (together with P and K), for plant growth (Osorio, 2007; Cordero et al., 2008). Biological nitrogen fixation done only by bacteria is an environmentally and economically viable alternative, to supply the inclusion of N sources of synthetic origin, because molecular nitrogen (N₂) is the only inexhaustible reserve in the biosphere.

In the rhizosphere of tropical and subtropical pastures there are N₂-fixing microorganisms; diverse studies have proven that there is high specificity of the isolations obtained from different grass roots, with regards to the diazotrophic species found, that is, root exudates, soil pH and other characteristics of the plant species and the soil could influence the presence of some genera of biological fixatives (Reis Junior et al., 2004; Radwan et al., 2005). Studies for more than 20 years indicate that the bacteria from the Azospirillum genus have special affinity for grass roots (Brasil et al., 2005), as in the case of the pastures that respond with increases in their growth and yield when they are inoculated with Azospirillum spp. The secretion of plant growth-promoting substances (such as auxins, gibberellins and cytoquinines) by Azospirillum, seems to be partially involved in this effect (Reis Junior et al., Radwan et al., 2005; Kuss et al., 2007).

Taking into consideration that in Colombia there is not any record of the evaluation of populations of this bacterium associated to pastures until now, the objective of this work was to begin the study of native Azospirillum strains in Guinea grass in Valle del Cesar, as the first step towards the search for isolations with biofertilizer potential.

MATERIALS AND METHODS

Determination of the effect of the rainy and dry seasons on the Azospirillum populations in Guinea grass. Samples were taken from the rhizospheric soil and the whole Guinea grass plant in the rainy season in September, 2006 (158,6 mm rainfall and 28,1ºC temperature) and in the dry season in February, 2007 (0 mm rainfall and 31ºC temperature). Both selected sites show contrasting conditions regarding soil management. At the E.S. Motilonia agronomic practices are developed such as the incorporation of organic matter to the soil, planting planning and rational grazing; at the Fernambuco farm overgrazing and intensive laboring of the soil are practiced.

The rhizospheric soil was obtained by removing it from the roots, from which dilutions up to 10^-6 were made in saline solution at 0,85% of NaCl. Of each dilution 0,1 mL were inoculated in three vials of NFb and LGI semisolid medium, for the recount of Azospirillum spp. and Azospirillum amazonense, respectively (Döbereiner et al., 1995). The pasture roots were washed and cut into 1 cm-fragments and were disinfected with 70º alcohol for 1 min; they were later submerged in sodium hypochlorite at 2% for 2 min and were rinsed twice successively in sterile water. One gram of these roots was weighed and they were macerated in 9 mL of sterile saline solution (0,85% NaCl). The leaves and stems were washed with distilled water and dried. One gram of this plant material was weighed cut in 2 cm-fragments, they were disinfected with 70º alcohol and macerated in 9 mL sterile saline solution (0,85% NaCl). Seriated dilutions were made up to 10^-6 from these macerates and they were inoculated in the same way as the rhizospheric soil sample. The inoculated vials were incubated for 5 days at 32ºC until the formation of a subsurface film as indicator of positive growth.

The population was quantified through the Most Likely Number method for three tubes, applying McCrady's table (Döbereiner et al., 1995). Three repetitions were made and a variance analysis was done according to the experimental design in divided plots with 2 (climatic seasons) x 2 (agronomic managements) x 3 (analyzed samples) factorial arrangement; Duncan's comparison test at 5% probability was used, with the statistical program SAS version 9.1.3.

Isolation and phenotypical identification of Azospirillum spp. The vials of the higher dilutions with positive growth were selected and replicated in a new semisolid medium. From this film the isolation was done in Congo Red agar dishes to select scarlet red colonies (Rodríguez-Cáceres, 1982) and in Batata agar to select pink, small and structured colonies, after a week of incubation at 32ºC (Döbereiner et al., 1995). The bacterial cells were observed under a microscope with 1000x magnification and a smear was done in distilled water to describe the shape and movement of the cells. The characteristics of spiral mobility, growth in Batata and Congo Red agar, were compared to the reference strains Azospirillum brasilense (Sp7) and Azospirillum lipoferum (Y2). Their similarities were estimated through the JACCARD coefficient and were grouped by the analysis between groups to make a dendrogram, using the program SPSS 10.0 for Windows.

Selection of the Azospirillum spp. strains with biofertilizer potential. The presumptive isolations for the Azospirillum genus were evaluated according to the activity of the nitrogenase enzyme and the production of plant growth-promoting indolic substances. The strains were replicated in Congo Red agar and cellular suspensions were obtained adjusted to a cell concentration of 1*10⁷ UFC/mL; 30 mL of 2% DYGS broth were inoculated and it was taken to incubation during 48 hours at 32ºC and 120 rpm. The obtained biomass was centrifuged at 8 000 rpm during 10 minutes. The supernatant was discarded and the cells were suspended in 30 mL of sterile phosphate buffer 0,06M and pH 7.0. From this cell suspension 100 µL were taken and inoculated in 50 mL of NFb broth without Bromothymol blue and vitamin solution, with a concentration of 200 µM of tryptophan as precursor of indoleacetic acid and 0,2 g/L of NH₄Cl as source of N. It was incubated in the dark during 48 hours, 120 rpm and 32ºC (Radwan et al., 2004). Ten mL of the culture were centrifuged at 8 000 rpm for 10 minutes and 2 mL of the supernatant were taken, to which 8 mL of Salkowsky reagent were added. The sample absorbance reading was made at 535 nm in spectrophotometer Spectronic 601 Milton Roy. The concentration of indolic compounds was calculated using the equation Y=0,0033X-0,0311 (R²=0,9855) where, Y=Absorbance of the sample at 535 nm and X=µM of idoleacetic acid (IAA). This equation was obtained with the lineal regression of the calibration curve in different concentrations of IAA (25, 50, 100, 150, 200, 250 and 300µM) with the addition of the Salkowsky reagent (Kuss et al., 2007).

For quantifying the in vitro biological nitrogen fixation the acetylene reduction method was used. The isolations were inoculated in 10 mL-flasks with 3 mL of semisolid medium (NFb or LGI) and were incubated for 24 hours at 32ºC. Afterwards, the stopper was replaced by a rubber one, sealing it hermetically, 10% of its atmosphere was substituted by acetylene and it was incubated during 1 hour at 32ºC. The ethylene was measured injecting 1 mL of the atmosphere of the culture flask in a Perkin Elmer gas chromatograph with flame ionization detector and a Poropak column N 200/300 mesh of 6 feet and 3 mm diameter. The acetylene reduction activity for each isolate was calculated according to the height of the ethylene peak in the chromatogram, extrapolating in the equation Log₁₀Y = 0,808 Log₁₀X + 9,7 (R² = 0,997), obtained from the lineal regression of a calibration curve in ethylene concentrations of 1, 1:10, 1:10, 1:1000 and 1:10 000, under the same conditions as the chromatograph (Corpoica, 2006). The production of indolic compounds and the biological nitrogen fixation were compared, and their similarities were estimated according to the Euclidian distance. Afterwards, they were grouped by the median method and graphically represented in a dendrogram, using the program SPSS 10.0 for Windows.

Identification of the Azosporillum spp. isolations by analyzing the gen 16S rrna. The selected strains were cultivated in LB broth for 14 hours at 150 rpm and 30ºC; 25 µL were taken and kept in thermostated bath during 15 minutes. Afterwards, it was centrifuged at 13 000 rpm for 1 min. The supernatant was discarded and the precipitate was re-suspended in 1 mL of sterile Milli-Q water. The amplification of the 16S rRNA region was done in 25 µL of a Buffer reaction mixture (1x), MgSO₄ (1,5 mM), dNTP (0,2 mM), Primer 27F (Edwards et al., 1989) and 1492R (0,2 mM) (Weisburg et al., 1991), 0,25 µL of Taq DNA Polimerase and 1,5 µL of DNA mold (isolation sample). The conditions of the thermocycler were: an initial cycle of denaturation (94ºC for 5 min), 50 denaturation cycles (94ºC for 30 sec), ringing (66ºC for 30 sec) and extension (72ºC for 1 min) and three cycles of final extension (72ºC for 10 min). The amplified fragments were observed by means of agarose gel electrophoresis at 0,8% in TAE conditioned with 0,5 µg/mL of ethidium bromide, to visualize the resulting strips. The amplified DNA was purified using QIAquick PCP Purification Kit Protocol of QIAquick® Spin Handbook of QUIAGEN, it was quantified in a NanoDrop Spectrophotometer ND-1000 and it was analyzed in the program ND-10 00 V.3.3.0 at a wavelength of 260 nm. The amplified product was sequenced in an ABI PRISM Genetic Analyzer, which makes the automatic sequentiation of DNA fragments marked with fluorochromes. The result was analyzed through BLAST (Basic Local Alignment Search Tool) in www.ncbi.nlm.nih.gov/BLAST/Blast.cgi.

RESULTS AND DISCUSSION

Effect of the season and agronomic management of Guinea grass on Azospirillum populations. According to the variance analysis, the recount of the Azospirillum population did not show significant differences according to the season and the agronomic management of Guinea grass. However, this population tended to be higher in the rainy season under an agroecological management of Guinea grass at the E.S. Motilonia as compared to the dry season in the two localities, which shows the influence of environmental factors on the bacterial population (table 1).

Brasil et al. (2005) reported a decrease of the number of diazotrophic bacteria in the root and soil of the grasses Axonopus purusii, Elyonurus muticus and Brachiaria humidicola, in the dry season. On the other hand, Reis Junior et al. (2004) obtained A. amazonense populations of 5,81 log units in roots of Brachiaria spp. in the winter, statistically different from 3,67 log units in the dry season in a tropical zone with dry season in winter. Yet, when they evaluated the A. amazonense population in a tropical zone without dry season they did not found significant differences.

Although a decrease in the Azospirillum spp. population in the dry season was observed, these bacteria can survive due to their physiological characteristics, because they show high capacity to change their metabolic activity, when the soil environmental conditions vary, mainly the availability of water, carbon and nitrogen and oxygen tension. Thus, the Azospirillum species can use ammonium, nitrate, nitrite, aminoacids and molecular nitrogen as nitrogen source (Steenhoudt and Vanderleyden, 2000) and, under unfavorable conditions, such as drought and nutrient limitation, they can experience certain morphological and biochemical changes to form structures similar to cysts which allow them to survive under adverse physical conditions, particularly desiccation (Sadasivan and Neyra, cited by Joe et al., 2009).

On the other hand, when the Azospirillum spp. populations were compared in the rhizospheric soil, roots and leaves of Guinea grass in NFb and LGI medium, significant differences were recorded (P<0,05) with regards to these three studied samples. The recount of cells in the leaves was observed to show the highest population value, followed by the roots and rhizospheric soil (table 2).

These results are similar to the ones reported by Brasil et al. (2005) in three grasses, who observed a higher number of bacteria in the leaves, in the NFb medium, and statistically similar values in LGI medium. The presence of bacteria in all the structures of the plant has been described in diverse studies, because although it is considered a rhizobacteria and colonizes mainly the elongation zone of the root hairs, it has also been found within the roots and the aerial part of several plant species (Brasil et al., 2005; Brasil et al., 2006; Kuss, 2006). Probably all Azospirillum species can develop this colonization mechanism and that is why they have been found in higher number colonizing the roots, located in the mucigel or penetrating the cortical cells of the root of different plants. Yet, the Azospirillum species have specific mechanisms to interact with the roots and colonize their interior, and even penetrate progressively the leaf tissue, through the cellulolitic and pecnitolitic enzymatic activity, invading the intercellular spaces and even being translocated through the vascular vessels towards the stem and leaves (Steenhoudt and Vanderleyden, 2000).

Phenotypical characterization of the obtained isolations. Sixteen isolations of the Azospirillum genus were selected according to the phenotypical characteristics, which were obtained from the rhizospheric soil, root and leaf samples of the pasture cultivated in the E.S. Motilonia of Corpoica and the Fernambuco farm (table 3).

Figure 1 shows the dendrogram according to the similarity matrix and an analysis between groups of the 16 isolations, with the reference strains SP7 (A. lipoferum) and Y2 (A. brasilense). Three groups were obtained with different distance values among them. Group I showed four isolations very close to strains Sp7 and Y2. Group II showed four isolations, but it was distant in 20% from group I. Finally, number III included eight isolations and formed an independent group with 25% distance from the first ones (I and II).

This first grouping allowed selecting eight isolations with a maximum of 20% distance with regards to strain Sp7 and Y2. The selected isolations were HGF3, SRGM4, SRGM3, SGRM2, RGM1, RGM3, RGM6 and HGF1, which were analyzed according to their nitrogen fixing activity and indoleacetic acid synthesis.

Selection of Azospirillum isolations with biofertilizer potential. Although it was originally suggested that the action mechanism of Azospirillum to promote plant growth was its ability to fix atmospheric nitrogen, at present alternative mechanisms have been proposed, such as the production of plant growth regulators or phytohormones (Carcaño-Montiel et al., 2006). For such reason, the selection of Azospirillum strains was done taking into consideration their nitrogen fixation activity and the production capacity of indolic compounds, among which statistical differences were observed for all the evaluated isolations (table 4).

Nitrogenase activity of the different isolations. The strain Sp7 (A. brasilense) recorded the highest value of acetylene reduction, with 845,6 nmol C₂H₄ h^-1 mL^-1. Regarding the native isolations, the strain SRGM2 showed an acetylene reduction activity of 10,72 nmol C₂H₄ h^-1 mL^-1, similar to the one reported for the isolations of A. lipoferum N7 and A. brasilense N8 with a nitrogenase activity of 6,6 and 7,95 nmol C₂H₄ h^-1 mL^-1, respectively (Mehnaz and Lazarovits, 2006). SRGM3, according to its acetylene reduction activity of 44,69 nmol C₂H₄ h^-1 mL^-1, was similar to the Azospirillum strains isolated from wild corn (Teocintle) with nitrogenase activity of up to 46,44 nmol C₂H₄ h^-1 mL^-1 (Carcaño-Montiel et al., 2006). On the other hand, SRGM4 was more efficient with a value of 123,70 nmol C₂H₄ h^-1 mL^-1, similar to the report for the strain Azospirillum doebereinerae isolated from Miscanthus sinensis cv. Giganteus, , which showed a value of 100 nmol C₂H₄ h^-1 mL^-1 for acetylene reduction (Eckert et al., 2001). The other isolations showed ARA values lower than 1 nmol C₂H₄ h^-1 mL^-1, similar to the A. brasilense strains Cd and Az39 with 0,162 and 0,1 nmol C₂H₄ h^-1 mL^-1, respectively (Perrig et al., 2007).

Although a variation was observed in the enzymatic activity of the biological nitrogen fixation, it could be proven that Azospirillum is an important genus associated to tropical grasses, capable of reducing the atmospheric nitrogen in ammonium under microaerophilic conditions, through the activity of the nitrogenase complex, which has an iron-molybdenum cofactor, which is the reduction site of N₂, and of the protein reductase dinitrogenase which transfer electrons from an electron donor to the protein nitrogenase (Steenhoudt and Vanderleyden, 2000).

Production of indolic hormones. When the production of indolic compounds was quantified according to the studied isolations, the strain RGM1 had a production of 45,18 µg/mL of IAA, followed by SRGM2 with 43,27 µg/mL of IAA, which represent the highest values and Sp7 (A. brasilense) which recorded 40,170 µg/mL of IAA. The isolations SRGM4, RGM3, RGM6, HGF1 and HGF3 produced between 38,12 and 23,15 µg/mL of IAA and the strain SRGM3 showed the lowest value, 3,46 µg/mL of IAA (table 4). The production of indolic compounds of SRGM3 is comparable to the ones reported by Perrig et al. (2007), who recorded 2,9 µg/ mL of IAA for the strain Az39 (recommended for corn and wheat in Argentina), which indicates that this a concentration that allows plant growth promotion, taking into consideration that these strains have been evaluated in grasses and selected for their optimum results. The other strains showed values between 23,15 and 45,18 µg/mL of IAA, which oscillate between those that have been recorded by different strains of A. amazonense isolated from roots of Brachiaria spp. in Brazil, with 6,13 to 19,27 µg/mL of IAA (Reis Junior et al., 2004), A. brasilense (Cd) and A. lipoferum (Br17), producing 58,71 and 63,49 µg/mL of IAA (Radwan et al., 2005), and isolations of Azospirillum spp. from corn plants in Mexico with values up to 49,65 µg/mL of IAA (Carcaño-Montiel et al., 2006).

Some studies have shown the quantitative variation of this biofertilizer activity in each evaluated strain, taking into consideration the genetic diversity they present regarding their ability to use some C and N sources, the presence of tryptophan as precursor of the indolic compounds, of their metabolic routes and the enzymatic activity regulated by different genes that can be expressed by the conditions of their natural environment (Radwan et al., 2004; Carcaño-Montiel et al., 2006), which varied according to the sampled zone, the plant stratum and the season.

In many reports, Azospirillum is considered the most important bacterial genus for crop growth or yield improvement in different countries (Perrig et al., 2007), especially for its high production of plant growth regulators (Radwan et al., 2004) such as auxins, several gibberellins and cytoquinines (Bashan et al. cited by Perrig et al., 2007), and 3-indoleacetic acid (IAA) is the main phytohormone produced by Azospirillum (Radwan et al., 2004). IAA is an auxin known for stimulating fast responses in growth, increase of cell elongation, cell division and their differentiation (Cassán et al., 2008).

Analysis of the groupings for the selection of the strains with biofertilizer potential. The dendrogram of the interaction of the medians of indolic compound production and the nitrogen fixing activity, showed two groups separated by more than 25% distance (fig. 2). The strains Sp7, SRGM3 and SRGM4 formed group 1, with 7% distance among its medians. In group II two subgroups were formed, where HGF1, HGF3, RGM1, RGM3 and RGM6 showed 2% distance and, in turn, this same subgroup (IIA) showed 7% distance from the strain SRGM2, forming subgroup IIB. This analysis of hierarchical classification allowed to select the isolations of group I as the strains of higher biofertilizer activity under in vitro conditions. For such reason, the strains SRGM3, SRGM4 and the strain SRGM2 were selected for their closeness to group I, where Sp7 (A. brasilense) was grouped.

Genetic identification of the Azospirillum spp. isolations. The primary analysis of similarity among the sequences of the obtained nucleotides and the ones deposited in the GenBank database using the BLAST analysis, showed that the isolations belong to the alpha subclass proteobacteria of the Rhodospirillacea family and are strongly related to the Azospirillum genus for presenting up to 94% similarity (table 5).

The strains SGRM2 and SGRM3 showed 93% similarity with A. lipoferum and SRGM4 94% similarity with A. brasilense.

Conclusions

No variations were found in the size of the Azospirillum population due to the season, which allows inferring that it tolerates hydric stress conditions.

From the soil, root, leaf and stem samples of Guinea grass 184 isolations were obtained, from which 16 were phenotypically characterized as belonging to the Azospirillum genus and eight were selected for their similarity with the studied genus.

Three promising strains were selected to be used as biofertilizer, because of their high nitrogen fixation and production of indoleacetic acid, they were molecularly identified and similarities were found with A. lipoferum and A. brasilense, according to the GenBank database.

Acknowledgements

To the Soil Microbiology Laboratory and the Experimental Station Motilonia of CORPOICA, for their collaboration and economic support through the project financed by the Ministry of Agriculture and Rural Development of the Republic of Colombia, the Colombian Federation of Livestock Raisers FEDEGAN and the Government of the Cesar Department.