Introduction

Frequently, chemical P fertilizers are applied to soils in order to meet P demands from crops, but they become rapidly transformed into immobile forms through precipitation reactions with aluminum or iron in acidic soils, and with calcium in normal or calcareous soils (Hao et al. 2002; Gyaneshwar et al. 2002). Under these conditions, the availability of soluble P-forms for plant growth becomes strongly limited. As an alternative to P fertilization, many soil microorganisms that play a key role in soil P dynamics can increase P uptake by plants. Inorganic P forms can be solubilized by phosphate-solubilizing bacteria (PSB), which excrete organic acids that dissolve phosphate minerals and/or chelate cationic partners of the PO4 3− anion,, thus releasing P into the solution (Vyas and Gulati 2009). Although these bacteria can be used as plant growth promoters, usually their amounts in soil are not enough to compete with other well-established bacteria in the rhizosphere, making it necessary to increase their levels by plant inoculation with adequate strains (Igual et al. 2001). In this sense, a pre-requisite for introducing these beneficial bacteria in the environment is, together with the ability to promote plant growth, to be adapted to the soil conditions and successfully compete with the soil microflora (Taurian et al. 2010). In this regard, the isolation and characterization of native bacteria could provide adequate strains for crop inoculation in particular soils where these microorganisms are well adapted.

The Salado River Basin is a typical breeding area for beef production in which perennial grasses predominate over other plant species, a common scenario in natural temperate regions. Thus, either as a consequence of land use or selective grazing, a decline or even absence of native perennial legumes is evident in this region (Montes 1988). In this way, the promotion of native or naturalized legumes present in these grasslands is an alternative to increase forage quality and yield. Lotus tenuis (=L. glaber Mill.), a perennial legume of European origin that was naturalized in the Salado River Basin, is highly valuable in this region because of its contribution to the forage offer and its positive influence on growth of associated species (Díaz et al. 2005). However, the successful implantation of this legume in the grassland community is usually limited by its low growth-rate during early developmental stages, which strongly decreases its ability to compete with weeds (Blumenthal and McGraw 1999). As a consequence, cultural practices that improve L. tenuis implantation are expected to increase forage yield and quality. Plant growth is restricted by several adverse factors in the Salado River Basin, such as alternating periods of water deficit and excess, as well as saline and alkaline soils with low nutrient content (Garcia and Mendoza 2008). In particular, soils in this region exhibit low P availability for plant growth, which seriously affects forage quality and productivity (Escudero and Mendoza 2005; Garcia and Mendoza 2008). In this way, inoculation of L. tenuis with selected PSB native from soils in this region is expected to improve P acquisition, thus improving growth of this legume and associated grasses. This approach could then be an affordable and sustainable practice to improve forage production in the Salado River Basin. In this regard, it should be kept in mind that in addition to being an essential nutrient for all plant species, P is particularly important for legumes due to the high ATP demand of the symbiotic nitrogen fixation process established with soil rhizobia (Robson et al. 1981). Therefore, this is an additional reason why the consequences of inoculating L. tenuis plants with PSB in the Salado River Basin are worth to be explored.

In a previous study aimed to improve the quality of bacterial inoculants for L. tenuis in the Salado River Basin, we isolated and characterized PSB from the rhizosphere of this legume. Thus, 17 gram-negative PSB belonging to different genera such as Pantoea, Erwinia, Pseudomonas, Rhizobium and Enterobacter were identified. Among them, Pantoea eucalypti isolate M91 exhibited the highest efficiency in promoting growth of L. tenuis plants through P solubilization, as shown by experiments in culture media containing Ca3(PO4)2 (non-soluble) as the sole P source. Thus, this was the first report about plant growth promotion by P. eucalypti, as well as the first report about promotion of L. tenuis growth by inoculation with PSB (Castagno et al. 2011).

The aim of the present work was to evaluate the ability of P fertilization and PSB inoculation to promote growth of L. tenuis in typical soils of the Salado River Basin. For this purpose, aboveground biomass, P and N levels were evaluated in field-grown L. tenuis plants inoculated with P. eucalypti M91, either without fertilization or in combination with phosphorous fertilizers of different solubility levels. Phosphate rock and triple superphosphate were used as low and high solubility phosphate sources, respectively, both of which were applied at a moderate concentration (20 kg P ha−1). The impact of P fertilization and P. eucalypti inoculation on the symbiotic interactions between L. tenuis and native rhizobia bacteria and arbuscular mycorrhizal (AM) fungi was also evaluated.

Materials and methods

Soil features and site location

The experimental plots were established in the Chacra Experimental Integrada de Chascomús (Ministerio de Asuntos Agrarios de la Provincia de Buenos Aires - Instituto Nacional de Tecnología Agropecuaria) located in Chascomús, Argentina (35° 30′ S, 58° 30′ W) during the period 2010–2011. Local air-temperature and rainfall data collected over 75 years (1911–1986) were provided by the Centro de Investigaciones Biometeorológicas (CIBIOM-CONICET) and were used to compare the temperature and rainfall recorded during the experimental period at the Experimental Station. The experimental plot was located in a typical area devoted to cattle breeding for beef production, sustained by natural pastures dominated by perennial and seasonal grasses, which lack traditional legumes such as red and white clover (Trifolium pratensis and T. repens) and lucerne (Medicago sativa). The experiment was conducted in a soil classified as a typical Natraquoll (Wermbter and Ramallo 1980) and its main chemical properties were pH: 6.7, EC (electrical conductivity): 2.4 dS m−1, C (organic carbon): 2.02 %, N (total nitrogen): 0.23 %, OM (organic matter): 3.5 %, P-availability (Bray I): 11.6 ppm, ESP (exchangeable sodium percentage): 13.2 % in the 0.20 m upper A1 horizon, followed by a 0.23 m B2t horizon with pH: 9.2, EC: 1.7 dS m−1, C: 0.36 %, N: 0.04 %, OM: 0.62 %, P Bray: 3.5 ppm P, ESP: 51 %.

Inoculum preparation and seed inoculation

The phosphate-solubilizing bacterium P. eucalypti M91 used in this work was isolated from the rhizosphere of L. tenuis growing in a typical lowland of the Salado River Basin and further characterized by biochemical and molecular studies (Castagno et al. 2011). In order to prepare the inoculant, P. eucalypti M91 strain was grown in liquid TY (Sperry and Wilkins 1976) at 30 °C on a rotatory shaker (180 rpm). Bacterial cultures were started in 3 ml medium and after 24 h they were sub-cultured into 250 ml flasks containing 50 ml of fresh culture medium. After an additional 24 h-incubation period, cultures were inoculated into 2 L flasks containing 500 ml of fresh medium and further incubated for 48 h. Then, this culture (1 × 109 bacteria /ml) was half-diluted with sterile water and used for seed inoculation, by spreading the bacterial suspension over the sown line..

Experimental design

The experimental area was sprayed with glyphosate twice, 40 days before sowing and a few days after sowing, in order to eliminate weeds. Additionally, Flumetsulam (Preside®) was applied after sowing and before emergence to eliminate broadleaf plants. Before sowing, the land was leveled by a disc harrow to homogenize sowing depth. After leveling, L. tenuis cv. Esmeralda was sown on August 24th (winter) in plots 9.0 m long and 1.5 m wide at a rate of 17 kg ha−1. Plots were fertilized with either triple superphosphate (TSP, 20 %P) or phosphate rock (PR, 13 %P) from Gafsa (Tunisia) at an equivalent rate of 20 kg P ha−1, or were not fertilized (−P). Two weeks after sowing, plots were divided in two subplots (4.5 × 1.5 m) and (PSB) P. eucalypti M91 was inoculated in solution by irrigating one of the two subplots of fertilized and non-fertilized plots. The experimental model was a randomized complete block trial in a split-plot design where the main plots corresponded to the P-sources (PR, TSP and -P) and the subplots to the inoculation treatment (inoculated and control (non-inoculated)). Treatments were replicated four times and the whole experiment thus consisted of 24 subplots.

Re-isolation of P eucalypti M91 from L. tenuis plants grown under controlled conditions

With the aim of assessing whether P. eucalypti M91 is able to endophytically colonize L. tenuis plants, seedlings derived from surface-disinfected seeds were inoculated with the above-mentioned strain and cultivated under controlled conditions, according to the methodology described by Sannazzaro et al. (2011). Endophytic colonization of L. tenuis plants by P. eucalypti M91 was evaluated by epifluorescence microscopy. For this purpose, leaves from inoculated and non-inoculated plants were surface-disinfected with 70 % (v/v) ethanol for 2 min and afterwards were immersed in a solution of acridine orange (0.1 %, pH 7.0) for 3 min, which allows visualization of both bacterial cells and anatomical structures of the leaf surface (Monier and Lindow 2004). Leaves were then air-dried and placed on top of a molten water-agar layer (1 %) previously spread on a microscope slide, to ensure a flat surface for microscopic observation. Samples were analyzed with a Nikon Eclipse E600 microscope equipped with a 450 to 490 nm excitation filter and a 520 to 560 nm emission filter. Images were acquired with a Nikon DS-Qi1Mc video camera.

Re-isolation and identification of P. eucalypti M91 from L. tenuis plants

In order to confirm the endophytic colonization of L. tenuis plants by P. eucalypti M91 under field conditions, endophytic PSB were isolated from roots and shoots at the second harvest. These organs were obtained from inoculated and non-inoculated plants and were surface disinfected as previously described. Plant material was then homogenized in 0.85 % NaCl and serial dilutions of the extracts thus obtained were plated on NBRIP agar medium (Nautiyal 1999) and incubated 48–72 h at 30 °C. Colonies showing a clear halo of phosphate solubilization were used to obtain DNA by the method described by Sannazzaro et al. (2011). BOX and REP-PCR patterns of the strains thus isolated were obtained and compared to those of P. eucalypti M91.

Aboveground biomass, plant and soil analyses

Three harvests were performed at different time points during the 261-day-period of the experiment: December 14th 2010, February 28th 2011 and May 11th 2011. Samples were taken from a 0.25-m2 area randomly chosen within the central rows of each sub-plot. Aboveground plant material was collected clipping 0.03 m above the soil surface. After each harvest, the experimental area was mown to 0.03 m and the aboveground material was removed from the plot.

Additionally, three plants were withdrawn from each subplot along with their rhizospheric soil. Roots were separated from the soil by washing in deionized water, and soil was used to analyze pH (1:2.5 suspension in water), EC, available P (Kurtz and Bray 1945), C (Richter and von Wistinghausen 1981), and total N (Bremmer and Mulvaney 1982).

Aboveground plant material obtained as described above was dried at 70 °C for 48 h and weighed. Yield was then calculated as plant biomass (kg) per hectare. P and N levels were determined by the molybdovanadophosphoric acid and Kjeldahl methods, respectively (Jackson 1964). For P determination, dry aboveground-plant material was previously digested in a nitric-perchloric acid mixture (3:2). For N determination, plant material was previously digested in sulphuric acid.

AM colonization and nodulation of L. tenuis roots by rhizobia

Mycorrhizal root colonization was measured in plants collected from each subplot, by using fresh roots cleared in 10 % KOH for 12 min at 90 °C and stained in 0.05 % lactic acid–glycerol Trypan blue (Phillips and Hayman 1970). Three plants per plot and twenty-five root segments per plant were examined under a microscope (×200). The fraction of root length colonized (MC), and the root length containing arbuscules (AC), vesicles (VC), and hyphae only (HO) was determined following McGonigle et al. (1990), and the entry points per mm of colonized roots was measured by the method of Amijee et al. (1989). The nodules formed on L. tenuis roots by native rhizobial populations of the experimental field were counted under a binocular stereomicroscope (×7.5).

Determination of AM fungal spore density in rhizospheric soil

Rhizopheric soil samples (250 g) were taken from each subplot and kept in plastic bags at 4 °C until used for the quantification of AM-fungal-spore density. AM fungal spores were isolated from soil by a modification of the sucrose gradient centrifugation technique (Daniels and Skipper 1982) and counted under ×35 magnification using a dissecting microscope. Spore density was expressed as the number of spores per 100 g of dry soil.

Statistical analyses

Data sets were tested for normality and heterogeneity of variance (heteroscedasticity) with standard methods. Non-normal data were appropriately transformed to compare treatment means. Analyses of variance (split-plot design) were used to test the equality of treatment means. Mean separation was performed by the LSD test. Treatment effects from data that could not be normalized by any of the common standard transformations were tested with the Kruskal-Wallis non-parametric test. Statistix 9.0 software was used for statistical analyses.

Results

Climate and soil

The soil used for the field experiment represented a typical condition of the Salado River Basin area, with a top horizon close to neutrality followed by a B2t horizon very poor in nutrients and OM, strong alkalinity with 51 % of ESP and 45 % of clay plus fine silt (Wermbter and Ramallo 1980). The drainage capacity was limited, and the soil was not suitable for crop production. During the experimental period (August–May), mean air temperature was 16.7 °C and accumulated rainfall was 524 mm. Air temperature was quite similar to the historical temperature recorded at the same location for 75 years (1911–1986), but the accumulated rainfall was 35 % lower (Fig. S1). Excepting for September, January and February, monthly rainfall was markedly lower before the first (October–December) and the third harvest (March–April), as compared to the historical recorded rainfall.

Effect of P fertilization and PSB inoculation on the aboveground biomass

P. eucalypti M91 inoculation exerted a highly significant effect on dry matter yield at the first harvest (P = 0.0001, Tables 1 and 2), as opposed to the second and third harvests (P = 0.6216 and P = 0.0989 respectively, Tables 1 and 2). In this way, inoculated plants exhibited a 68 % increase in dry matter yield at the first harvest, as compared to non-inoculated controls (Table 1). P fertilization had no effect on biomass yield at any harvest (Tables 1 and 2). Analysis of variance also revealed that the interaction between P fertilization and inoculation with PSB was not significant in terms of dry matter yield at any harvest (Table 2). The multiple comparison of results obtained at the first harvest showed a tendency to increase biomass production of inoculated plants, regardless of P addition and the P source used (Table 1). A significant increase in biomass production was obtained in inoculated, non-fertilized plants or fertilized with PR, as compared to their respective non-inoculated controls (Table 1).

Table 1 The effect of P fertilization and inoculation with P. eucalypti M91 on the aboveground dry yield, the concentration of P and N for each harvest in shoot tissue of L. tenuis
Full size table
Table 2 Split Plot ANOVA (F and P values) for the effect of P-source (main factor) and inoculation with P. eucalypti M91 (subplot factor) on the aboveground dry yield, the concentration of P and N, and the accumulated P and N in L. tenuis shoot tissue
Full size table

Inoculation of L. tenuis with P. eucalypti M91 significantly affected accumulated dry matter yield (P = 0.0104, Table 2), and P-source had no effect on accumulated dry yield (P = 0.0893, Table 2). A multiple comparison of accumulated biomass showed a similar tendency to that observed in the first harvest, evidenced as an increase in the biomass of inoculated plants, as compared to non-inoculated ones regardless of P fertilization (Fig. 1a).

Fig. 1
figure 1

Accumulated dry matter yield, P and N at the three harvests in shoot tissue of L. tenuis in response to P fertilization and inoculation with P. eucalypti M91. Accumulated dry matter, P and N, discriminated for each combination of P-source and inoculation treatment, are shown in panels (a), (b) and (c), respectively. Results are presented as means ± standard error. Different letters indicate significant differences between treatments at a level of P < 0.05 by the LSD test. Accumulated dry yield was log transformed for this analysis. –P, no fertilization; +PR, fertilization with phosphate rock; +TSP, fertilization with triple superphosphate. White bars, non-inoculated plants; grey bars, PSB-inoculated plants

Full size image

Effect of P-fertilization and PSB inoculation on shoot-P and N levels

In order to evaluate the impact of PSB inoculation and P fertilization on plant nutrition, P and N levels were analyzed in plant shoots. P concentration was significantly affected by PSB inoculation at the first and second harvests (P = 0.0012 and P < 0.0001, Tables 1 and 2). Thus, P concentration of PSB-inoculated plants was 15.4 and 13.3 % higher than non-inoculated plants, for the first and second harvests, respectively (Table 1).

A multiple comparison of the results obtained for the first harvest showed that P concentration of PR- and TSP-fertilized and inoculated plants increased by 7.7 and 14.3 %, as compared to their respective non-inoculated controls (Table 1). For this harvest, a similar tendency was observed for inoculated plants that were not fertilized, although differences in P concentration observed between inoculated and non-inoculated plants were not statistically significant (Table 1). At the second harvest, P concentration was higher in inoculated than in non-inoculated plants (Table 1). In this way, inoculation increased P concentration by 28.6 % in TSP-fertilized plants, 10.3 % in PR-fertilized plants and 9.4 % in non-fertilized plants (Table 1). At the third harvest, no differences in P concentration were detected between inoculated and non-inoculated plants, regardless the fertilization treatment (Table 1).

Fertilization had a significant effect on shoot-P concentration only at the first harvest (P = 0.0074, Table 2), mainly due to the 15.4 % increase in P-concentration of TSP-fertilized, as compared to PR-fertilized and non-fertilized plants, respectively. At the three harvests, accumulated P in shoot tissue was also higher for inoculated plants, as compared to non-inoculated ones (P = 0.0234, Table 2). P content was significantly higher in plants inoculated with PSB and fertilized with TSP than in all the other treatments (Fig. 1b).

PSB inoculation significantly affected shoot-N concentration at each of the three harvests, with P-values = 0.0054, 0.0078 and 0.0042, respectively (Table 2). Thus, N concentration of inoculated plants was 14.6, 10.0 and 13.1 % higher than non-inoculated controls for the first, second and third harvests, respectively (Table 1). Shoot-N concentration was also affected by fertilization, an effect that was evident at the second and third harvests (P = 0.0123 and 0.0272, Table 2). At the second harvest, both PR and TSP fertilization exerted a negative effect on shoot-N concentration, which decreased by 20.5 % in PR fertilized plants and 13.8 % in TSP-fertilized plants, as compared to non-fertilized ones (Table 1). At the third harvest, shoot-N concentration was decreased by 10 % in TSP-fertilized as compared to non-fertilized plants, while PR-fertilized plants exerted no effect on shoot-N concentration, as compared to non-fertilized plants (Table 1).

A multiple comparison of results obtained at the first harvest showed that shoot-N concentration of inoculated plants fertilized with PR or TSP increased by 15.9 and 25.7 % respectively, as compared to non-inoculated controls (Table 1). At the second harvest, N concentration of inoculated plants increased by 19.3 % in TSP-fertilized plants, as compared to their respective non-inoculated controls (Table 1). For both non-fertilized and PR-fertilized plants, inoculation had no effect on shoot-N concentration (Table 1). At the third harvest, all the inoculated plants showed similar N concentration, were they fertilized or not, with the exception of non-inoculated plants fertilized with TSP, which exhibited lower nitrogen concentration than all the other treatments (Table 1). PSB inoculation significantly affected N accumulation in shoots at the three harvests (P = 0.0090, Table 2). Accumulated N was significantly higher in plants inoculated with PSB and fertilized with TSP than in non-inoculated plants (either fertilized or non-fertilized) (Fig. 1c).

Effect of P fertilization and PSB inoculation on L. tenuis symbiosis with AM fungi and rhizobia

The impact of P fertilization and PSB inoculation on the symbiotic relationships between L. tenuis and native soil microorganisms was analyzed at the first and third harvests. Regarding the interaction with AM fungi, MC and AC indexes, as well as the number of entry points in roots were not affected by the P source and PSB inoculation at any of the two harvests (Table 3). MC and AC indexes were always high, ranging from 0.75 to 0.87 (MC) and 0.63 to 0.74 (AC) for the first harvest and from 0.64 to 0.79 (MC) and 0.54 to 0.77 (AC) for the third harvest. The number of entry points was 4.2 and 2.7 per mm of colonized root for the first and third harvest, respectively, and this parameter was not affected by P fertilization and PSB inoculation (Table 3).

Table 3 Split Plot ANOVA (F and P values) for the effect of P-source (main factor) and inoculation with P. eucalypti M91 (subplot factor) on the performance of L. tenuis symbionts
Full size table

Fertilization had no effect on AM spore density at the first harvest (P = 0.1247, Table 3), while it exerted a significant effect at the third harvest (P = 0.0018, Table 3). PSB inoculation affected AM spore density at both harvests (P = 0.0012 and 0.0002 respectively, Table 3). At the first harvest, inoculation caused a decrease in spore density in non-fertilized and TSP-fertilized plants by 60 and 67 % respectively, but exerted no effect on spore count in soils fertilized with PR (Fig. 2a). At the third harvest, PSB inoculation increased spore density in plants exposed to all the fertilization treatments, between 33 and 73 % for PR and TSP, respectively (Fig. 2b). In this case, there was a clear negative effect of TSP fertilization on spore density both for inoculated and non-inoculated treatments (Fig. 2b).

Fig. 2
figure 2

Performance of L. tenuis symbionts in response to P fertilization and inoculation with P. eucalypti M91. Spore density of AM fungi in rhizospheric soil (a and b) and the number of rhizobial nodules in L. tenuis roots (c and d) were analyzed for the first and third harvests for each combination of P fertilization and inoculation treatment. Spore density data were arcsin transformed for this analysis and plotted back to original data. Results are presented as means ± standard error. Different letters indicate significant differences between treatments (P < 0.05) according to the LSD test. –P, no fertilization; +PR, fertilization with phosphate rock; +TSP, fertilization with triple superphosphate. White bars, non-inoculated plants; grey bars, PSB-inoculated plants

Full size image

The number of nodules per gram of root fresh weight was not affected by fertilization (P = 0.0592) and PSB inoculation (P = 0.2116) at the first harvest (Table 3). At the third harvest, this parameter was significantly affected by P fertilization (P = 0.0140, Table 3). As opposed to fertilization, PSB inoculation had no effect on root nodulation by rhizobia (P = 0.8196, Table 3). At the first harvest, the number of nodules did not differ between fertilized and inoculated plants and control plants (Fig. 2c). At the third harvest, fertilization with PR and TSP significantly reduced root nodulation of inoculated plants, as compared to non-fertilized inoculated plants (Fig. 2d).

Effect of P fertilization and PSB inoculation on soil chemical properties

Fertilization with TSP significantly increased available soil-P at the first harvest (P = 0.0437, Table 3 and Fig. 3a) but fertilization with PR had no effect on available P compared to non-fertilized soil (Fig. 3a). As opposed to the first harvest, fertilization with PR or TSP had no effect on available soil-P levels at the third harvest. In this case, available soil-P levels were similar for all treatments (Fig. 3b), being close to 10 mg P kg−1 and thus approaching to available soil-P levels measured at the first harvest for non-fertilized treatments (Fig. 3a). Inoculation with PSB had no effect on available soil-P measured by the Bray I soil test at both harvests (Fig. 3a and b). Other soil properties evaluated (pH, CE, C, N) were not significantly affected by P-fertilization or PSB-inoculation at both harvests (data not shown).

Fig. 3
figure 3

Available soil-P concentration in response to P fertilization and inoculation with P. eucalypti M91. Available-P was analyzed in soil samples at the first (a) and third (b) harvests for each combination of P-source and inoculation treatment. Results are presented as means ± standard error. Different letters indicate significant differences between treatments (P < 0.05) according to the LSD test. –P, no fertilization; +PR, fertilization with phosphate rock; +TSP, fertilization with triple superphosphate. White bars, non-inoculated plants; grey bars, PSB-inoculated plants

Full size image

Ability of P. eucalypti M91 to colonize L. tenuis plants as an endophyte

P. eucalypti M91 was found to endophytically colonize roots and leaves of L. tenuis plants grown in vitro (Fig. 4a). In order to determine if this bacterial strain is also able to act as an endophyte under field conditions, phosphate-solubilizing endophytic bacteria were isolated from roots and shoots of non-fertilized plants (either inoculated or non-inoculated) at the second harvest. In this way, a total of 51 isolates were obtained (Table S1). Interestingly, PSB-inoculated plants differed from non-inoculated ones in terms of endophytic colonization, and a higher number of isolates was recovered from PSB-inoculated plants, as compared to non-inoculated controls. Moreover, no endophytic bacteria were recovered from leaves of non-inoculated plants, all the 11 isolates from these plants being derived from root samples (Table S1). On the contrary, when PSB inoculated plants were analyzed, endophytic bacteria were recovered not only from roots (24 isolates), but also from leaves (16 isolates) (Table S1).

Fig. 4
figure 4

Ability of P. eucalypti M91 to colonize L. tenuis plants as an endophyte. a Leaf-endophytic colonization of in vitro-grown plants was confirmed by acridine orange staining and epifluorescence microscopy (100×). P. eucalypti cells (white arrows) were detected in leaf intercellular spaces of inoculated plants. b Field-grown plants inoculated with P. eucalypti M91 were found by REP-PCR analysis to harbor this bacterial strain as an endophyte, both in leaves and roots. Profile analysis was carried out on a range of fragments comprised between 2027 and 125 bp, using Lambda DNA/EcoRI+HindIII as a molecular marker (lane 1 and 11). Some examples of isolates that showed REP profiles identical to strain M91 (lane 2) are indicated with arrows

Full size image

REP-PCR profiles were obtained in order to determine if P. eucalypti was present within the collection of endophytic bacteria obtained from field grown plants. This analysis revealed that four out of the 24 profiles obtained from PSB-inoculated roots, as well as two out of the 16 profiles obtained from PSB-inoculated leaves were identical to that of P. eucalypti M91, the bacterial strain used for field inoculation (Fig. 4b). A similar analysis, based in BOX-PCR profiles, provided identical results (data not shown). Thus, the results obtained in the present work clearly demonstrate that P. eucalypti M91 is able to endophytically colonize L. tenuis plants in the field.

Discussion

The present work aimed to evaluate the potential of PSB inoculation and different P fertilizers to promote growth of L. tenuis plants in soils typical of the Salado River Basin. For this purpose, the native PSB P. eucalypti M91, previously demonstrated to promote growth of L. tenuis plants in vitro, was evaluated as an inoculant under field conditions, either without further amendments or in combination with chemical P-fertilizers (PR and TSP).

Previous works demonstrated that in P-deficient soils, L. tenuis strongly responds to the addition of this nutrient, both in monocultures and as a member of grassland communities (Ginzo et al. 1982). In the present work, L. tenuis was fertilized with a moderate P dose, as compared to the higher rates (100–200 kg TSP/ha) commonly used in the region (Ginzo et al. 1982; Mendoza et al. 1984; Bailleres and Pirodi 2000). The idea behind using this moderate P dose was to minimize the risk that high levels of P fertilization masked the effects P. eucalypti M91 inoculation on L. tenuis, thus enhancing the possibility of detecting growth-promoting effects of the above mentioned bacterium.

At the first harvest, inoculation with P. eucalypti M91 had a positive impact on L. tenuis biomass production. By contrast, inoculation did not improve biomass production at the second and third harvests, as compared to non-inoculated treatments. Even though P. eucalypti M91 inoculation did not to promote L. tenuis growth during the whole experiment, the promoting effects detected at the first harvest are important due to the fact that L. tenuis exhibits a low growth rate during initial growth stages, which significantly decreases its ability to compete with aggressive broad leaf- or grass weeds in the grassland community (Blumenthal and McGraw 1999).

It is interesting to note that accumulated biomass production of L. tenuis plants was increased by inoculation with P. eucalypti M91 in a similar degree to TSP and PR fertilization. Thus, PSB inoculation could represent a low-cost alternative for the improvement of L. tenuis yield without the disadvantages associated to high rates of P fertilization (Hao et al. 2002; Gyaneshwar et al. 2002).

It’s also worth to note that the application of P fertilizers (PR or TSP) without PSB had no significant effect on accumulated biomass production of L. tenuis. This could be the result of the moderate P concentration used in the present work to enhance the effect of PSB inoculation. Also, the drought conditions that prevailed during the experimental period, evidenced by an accumulated rainfall 35 % lower than the historical record (1911–1986), probably contributed to P immobilization after the application of fertilizers, thus leading to similar levels of available P in all the plots. In addition, the neutral and alkaline pH of the two first horizons may have contributed to decrease the solubility of fertilizers, especially PR. This situation is consistent with the values of available P found in soils of the experimental plots at the first and third harvests.

In regard to biomass production at the first harvest, PSB inoculation had a positive impact on shoot P concentration, as demonstrated by the significant increase in shoot P levels of inoculated plants, as compared to non-inoculated controls. This positive effect of PSB inoculation on shoot P concentration persisted up to the second harvest. At the first harvest, shoot P concentration was also significantly affected by fertilization. Moreover, the combination of P. eucalypti M91 inoculation and TSP fertilization significantly increased P accumulation in shoot tissues, as compared to other treatments. These results suggest that inoculation with P. eucalypti M91 increased available-P levels as a result of its solubilizing activity, thus allowing a more efficient use of P derived either from chemical fertilizers or from P compounds already present in the soil. Similar results have been reported for field trials in other crops such as rice, soybean, sunflower, maize and wheat, where an increase in biomass production and P content was observed after PSB-inoculation (Costa et al. 2013; Rana et al. 2012; Rodríguez and Fraga 1999). In a previous study by our group (Castagno et al. 2011), P. eucalypti M91 was found to promote L. tenuis growth under controlled conditions with tricalcium phosphate as the sole P source. Moreover, the abovementioned work demonstrated that the ability of this bacterial strain to solubilize tricalcium phosphate is related to gluconic and ketogluconic acid production (Castagno et al. 2011). P plant nutrition can also be improved by inoculation with bacteria that produce phytase, an enzyme capable of mineralizing organic P forms, which are highly abundant in soils (Lei et al. 2007; Pontoppidan et al. 2012). In this regard, P. eucalypti M91 was also found to be able to mineralize organic P forms such as Na and Ca phytate (unpublished results). On the basis of the information discussed in preceding paragraphs, the increase in P concentration and content detected in shoots of L. tenuis plants inoculated with P. eucalypti M91 could be attributed to the ability of this strain to solubilize both organic and inorganic P forms, thus increasing the availability of this nutrient in the soil.

In addition to the abovementioned effects on P nutrition, PSB inoculation also led to a significant increase in shoot-N concentration and accumulated N. On the contrary, the application of P fertilizers negatively affected shoot-N concentration, as shown by the comparison of PR- and TSP-treated plants with non-fertilized ones, particularly at the second and third harvests. Nevertheless, the highest accumulation of this nutrient in shoots was achieved by the combination of P. eucalypti M91 inoculation and TSP fertilization, a treatment that significantly increased accumulated N in shoots as compared to non-inoculated plants, regardless of fertilization. In this regard, increased N levels in shoot tissue of legumes are often related to the establishment of a N-fixing symbiotic interaction with native soil rhizobia, or in response to the inoculation with specific rhizobial strains (Zahran 1999; Sessitsch et al. 2002). The hereby-reported increase in shoot-N concentration of L. tenuis plants inoculated with P. eucalypti M91 cannot be attributed to a stimulatory effect of P. eucalypti M91 on nodulation with native rhizobia, at early stages of plant growth. In this regard, nodule numbers at the first harvest showed no significant variations between different fertilization and inoculation treatments. However, P. eucalypti M91 was able to stimulate nodulation at later growth stages in plants cultivated in absence of P-fertilizers. This view is supported by the significant increase in the number of nodules detected in non-fertilized plants inoculated with P. eucalypti M91, as compared to inoculated plants fertilized with PR or TSP. Legumes are known to regulate nodulation as a function of soil N levels and plant N content (Stougaard 2000). Thus, it wouldn’t be surprising that the increase in N concentration of TSP-fertilized and inoculated plants can act as a signal to down-regulate nodulation, which could in turn explain the decrease in nodule numbers detected at the third harvest in these plants

The increase in N content of plants inoculated with P. eucalypti M91 and fertilized with TSP, can also have resulted from an improvement of the nitrogen-fixing capacity of native rhizobia due to an increase in P availability for the energetic demands associated to nitrogenase activity. In this sense, some authors have demonstrated that biological N-fixation is highly dependent on the amount of available P in soil (Almeida et al. 2000). Therefore, part of the P solubilized by P. eucalypti M91 and assimilated by L. tenuis, could be transported to the nodules and used by the native rhizobia inside them, for metabolic processes that consume high levels of energy, (Sa and Israel 1991; Hernández et al. 2009). In addition, some Pantoea species are known to have nitrogenase activity and therefore increase N content in other plant species such as sugarcane (Loiret et al. 2004). Therefore, the possibility that P. eucalypti M91 is able to fix atmospheric N and thus contributed to increase N-content of inoculated L. tenuis plants cannot be discarded.

As a summary of the beneficial effects of the combination of P. eucalypti inoculation and TSP fertilization discussed so far, it is important to note that even though an increase in plant biomass production was evident only at the first harvest, a positive effect on P and N accumulation was evident for longer periods, being evident up to the second and third harvest for P and N, respectively.

Symbiosis with AM fungi significantly facilitates P acquisition by plants, and their contribution to plant growth has been well-studied (Smith and Read 2008). So, the impact of P fertilization and PSB inoculation on AM fungi colonization of L. tenuis roots was analyzed in the present work. L. tenuis roots were heavily colonized by AM fungi, as previously described for plants of this species naturally grown in grasslands of the Salado River region (Escudero and Mendoza 2005; Garcia and Mendoza 2007, 2008). Mycorrhizal colonization and the number of entry points were not affected by the combination of P-fertilization and inoculation with the native phosphate-solubilizing bacterium. These results indicate that both the growth of AM fungi within the roots and the formation of new colonization units were not affected either by the P dose used or by the inoculation with P. eucalypti M91. It is important to highlight that biofertilization with P. eucalypti M91 did not affect the development of the symbiotic relationship between L. tenuis and the native community of AM fungi in terms of the abovementioned parameters related to root colonization. It would be interesting, however, to investigate if under controlled conditions, P. eucalypti M91 affects L. tenuis-AM fungi symbiosis from a functional point of view, as well as to analyze if both groups of soil microorganisms can develop a synergistic interaction, as previously reported for other PSB and AM fungi by several authors (Artursson et al. 2006; Muthukumar and Udaiyan 2010). Unlike intraradical colonization, spore density showed a marked increase at the third harvest, which coincided with the beginning of autumn. This increase in spore density could be attributed to a seasonal effect that became evident after the summer, a period during which the plant and fungal communities were exposed to high temperature and dry periods. Seasonality has been previously described as one of the factors able to modify the community of AM fungi in the rhizosphere of L. tenuis grown in different areas of the Salado River Basin (Escudero and Mendoza 2005; Garcia and Mendoza 2007, 2008). Another important point is that spore density was affected by the addition of TSP, a highly soluble P source. Thus, the increase in P availability in soil observed at the first harvest may have inhibited the formation of new spores, as previously described by de Miranda and Harris (1994) and Smith and Read (2008).

The present work demonstrated that P. eucalypti M91, previously isolated from rhizospheric communities, is also able to endophytically colonize L. tenuis roots and leaves, both under controlled and field conditions. This finding is interesting, because the endophytic life-style could enhance the ability of this strain to survive in natural environments, once applied as an inoculant. Interestingly, under field conditions, plants inoculated with P. eucalypti M91 were also endophytically colonized by other PSB. In this regard, the presence of higher numbers of other endophytic PSB in P. eucalypti M91-inoculated plants, as compared to non-inoculated ones, suggests that inoculation with a single PSB strain somehow enhances plant colonization by other strains. The elucidation of the mechanism by which inoculation with a particular endophytic PSB leads to increased endophytic colonization by other PSB strains deserves further research. It is also worth to mention that results obtained in the present work demonstrate that specific PSB strains used as inoculants can be traced in the field, which could be a useful tool for the analysis of inoculation efficiency and the population dynamics of such bacteria in plant hosts, as well as in the environment.

The results of this study suggest that inoculation of L. tenuis with phosphate-solubilizing bacteria from soils of the Flooding Pampa, such as P. eucalypti M91 would be a sustainable option for grassland producers of this region, rather than chemical fertilizers, since it would allow a more efficient use of nutrients such as nitrogen and phosphorus.