Effect of Bacillus subtilis on phosphorus uptake by cucumber as affected by iron oxides and the solubility of the phosphorus source

In this work, we examined the effects of Bacillus subtilis strain QST713 by assessing plant P uptake from variably P compounds. The experiment performed involved three factors: (i) P source (KH2PO4 at 100 mg kg –1, and phosphate rock [PR] at 100 or 200 mg kg–1); (ii) plant inoculation with QST713 (inoculated and non-inoculated); and (iii) Fe oxide (ferrihydrite) in the growth medium (0 or 300 mg kg–1 concentration of citrate–ascorbate-extractable Fe). Ferrihydrite decreased dry matter yield in plants by more than 50 %. Inoculation with QST713 increased plant growth, and total accumulation of P by plants and P uptake, estimated as the total P in aerial parts and roots minus P present in seeds. Overall, QST713 increased P uptake by 40 %, the effect being independent of the presence of ferrihydrite and P source. The increased P uptake observed can be ascribed to solubilization of P and to increased root growth. Therefore, QST713 improves P nutrition in plants grown on media with a high P adsorption capacity irrespective of the solubility of the P compound.


Introduction
Although agricultural soils usually contain substantial amounts of total P, P deficiency is a widespread problem in agriculture (Holford 1997, Wang et al. 2011) because most of total P is present as low plant-available forms owing to its reactions in soil (Delgado andScalenghe 2008, Ryan et al. 2012).Among such reactions, precipitation of metal phosphates and P adsorption onto Al-and Fe-oxides, are known to play a crucial role in the dynamics and availability of this nutrient in soils (Schwertmann and Taylor 1989, Guzman et al. 1994, Saavedra and Delgado 2005).A high P fixing capacity in soils such as those with high contents of Fe oxides decreases the efficiency of applied fertilizer in increasing P availability in soil (Ryan et al. 2012).
Today's agriculture relies on regular inputs of phosphate fertilizer derived from phosphate rock, a non-renewable, strategic resource (Whiters et al. 2014).The predicted future scarcity of which is raising fertilizer prices and emerging as a critical issue that may affect global food security in the near future (Gunther 2005, Cordell et al. 2009, Cordell and Neset 2014, Amundson et al. 2015).This requires an efficiently management of available P resources (Simpson et al. 2011, Withers et al. 2014), through a better understanding of the biogeochemical cycle of P, the design of more rational fertilization schemes, as well as the use of more efficient biological resources such as crops with an increased capacity to use it (Kochian 2012, Heppell et al. 2015).The manipulation of the rhizosphere by inoculation with microorganisms capable of improving P nutrition of crops can also contribute to the efficient management of P resources (Patel et al. 2010, Wang et al. 2011, García-López et al. 2015).
Rhizosphere microbial communities can improve uptake of some nutrients such as P and hence it may contribute to overcome nutritional deficiencies (Aseri et al. 2008, Zuo and Zhang 2011, de Santiago et al. 2013).Some P-solubilizing microorganisms have been recommended for use as inoculants by virtue of their ability to increase P availability to plants by mobilizing P from scarcely available forms such as insoluble Ca phosphates or organic P forms naturally present in or applied to soil (Richardson 1994, 2001, Illmer et al. 1995, Whitelaw et al. 1999, Oliveira et al. 2009).Rhizospheric microorganisms can have additional benefits such as plant growth promotion (PGPR) or facilitating biocontrol of soil-borne diseases (Vassilev et al. 2006, de Santiago et al. 2011).The benefits of microbial inoculants on crops can provide a basis for their use in cost-effective integrated agricultural practices.Some commercial strains of Bacillus subtilis have proved effective PGPR and biocontrol agents (Yuan et al. 2012); others, have been found to facilitate solubilization of P compounds (Orhan et al. 2006, Mena-Violante andOlalde-Portugal 2007).

Manuscript received May 2016
Phosphorus availability to plants is related to the solubility of metal phosphates present in soil -Ca phosphates in alkaline soils, mainly-and to P desorption from sorbent surfaces such as Fe oxides.The action of P-mobilizing microorganisms may be affected by the sorption capacity of the growing medium and the solubility of P forms in it.The influence of the solubility of P compounds may be better understood by elucidating the role of scarcely soluble Ca phosphates in calcareous media in supplying P to inoculated plants.These P compounds are major residual sources of P unavailable to plants in calcareous soils.Although the literature on P-mobilizing organisms is growing steadily, there is still no unambiguous information about the main P forms being mobilized and how their mobilization is related to soil properties affecting P dynamics in soil, and particularly to Fe oxides.Microbial mechanisms enhancing Fe uptake by plants through dissolution of Fe minerals, such as release of complexing agents, acidification, or Fe(III) reduction (Lemanceau et al. 2009), may contribute to the uptake of P bound to Fe compounds by plants (Ghosh et al. 2015).Thus, an increased Fe uptake by inoculated plants with P-mobilizing microorganisms may reveal that improved P nutrition is ascribed to the release of P bound to Fe compounds.The main objective of this research was to understand the effect of inoculation with PGPR Bacillus subtilis QST713 on P uptake by plants and whether the effect is influenced by the presence of Fe oxides (ferrihydrite) in the growing medium or the solubility of the P source applied.QST713 was selected because it is a commercial strain of B. subtilis effective in promoting plant growth and in the biocontrol of some relevant soil-borne diseases (Lahlali et al. 2011).Thus, evidence of its effectiveness as a P-mobilizing inoculant under adverse conditions (e.g., in media with scarcely soluble dominant P forms and a high P adsorption capacity) is bound to support its use in costeffective integrated management practices contributing to agricultural sustainability in the near future.

Experimental design
An experiment with cucumber plants (Cucumis sativus L. cv Tropico) grown in an artificial medium was performed.This type of medium was used to assess P mobilization from specific P forms because the wide range of compounds present in agricultural soils precludes identifying the origin of P absorbed by plants.The growing medium was a mixture of siliceous (70%) and calcareous (30%) sand.Before use, both sands were sieved between 0.5 and 1 mm to facilitate aeration and increase hydraulic conductivity.Siliceous sand was washed several times with diluted Na 2 CO 3 to disperse and remove clay minerals and Fe oxides, whereas calcareous sand was washed with deionized water.The final pH of the growing medium, measured in water at a 1:2.5 growing medium/solution ratio, was 8.5 -a typical value for the calcareous soil the growing medium was intended to mimic.
The plants were inoculated with B. subtilis strain QST 713 (Serenade ASO, Bayer Cropscience) by adding 2 × 10 7 colony forming units (CFU) per kilogram of growing medium.To this end, a volume of 20 ml of an aqueous suspension containing 4 10 8 CFU l -1 was applied to the soil surface divided over four points around each plant after transplanting.Non-inoculated plants received an identical volume of de-ionized water that was applied in the same manner.The amount of nutrients supplied by the inoculant was negligible.Also, using Hoagland solution, which contains high nutrient concentrations, offset any potential differences in nutrient supply by effect of its rapidly replacing the solution in the sandy growing media after irrigation.

Plant analysis
After 28 days of growth, plants were harvested, and the shoots and roots of each plant separated and dried at 65 °C to constant weight in a forced-air oven.Then, dry matter (DM) in shoots and roots was measured, and P and Fe concentration in both organs were determined.To this end, dried plant material was ground to pass through a 1-mm sieve and an aliquot of 0.25 g was placed in a porcelain crucible for mineralization by dry combustion in a furnace at 550 °C for 8 h.After combustion, the ash was dissolved in 10 ml of HCl (1 mol l -1 ) and the resulting solution heated at 100°C for 15 min to recover all nutrients.Phosphorus in the digest was determined by using the colorimetric procedure of Murphy and Riley (1962), and Fe by atomic absorption spectrometry on a Unicam Solaar M instrument from Thermo Fisher Scientific (Madrid, Spain).Certified material (1 sample in each combustion run) was also analysed in parallel to assess total recovery of nutrients present in the plant material.Total P uptake by plants was estimated as that present in plant tissues minus that in seeds.

Growing medium analysis
Alkaline phosphatase activity at the end of the experiment was determined in the rhizospheric growth medium according to Tabatabai and Bremner (1969), i.e., as the amount of PNP (p-nitrophenol) formed from pNPP (p-nitrophenyl phosphate).
Organic anions (OA) were extracted from rhizospheric soil by shaking an amount of 5 g of sample with 5 ml of NaOH (0.1 mol l -1 ) at 4 Hz during 1.5 h (Baziramakenga et al. 1995, Radersma andGrierson 2004).After shaking, suspensions were centrifuged at 10,000 g for 10 min and the supernatants were filtered through Whatman 42 filter paper.The filtrate from each NaOH extract was acidified to pH 2-3 with H 2 SO 4 (1 mol l -1 ) and the supernatant passed through a filter of 0.45 μm pore size.Organic anions were separated by high performance liquid chromatograph (HPLC) on an HPLC Varian ProStar 410 instrument furnished with a C18 column (Varian, 250 mm × 34.6 mm, 8 µm particle size), using isocratic elution with 98 % of H 2 SO 4 (5 mmol l -1 ) at pH 2 plus 2 % methanol at 0.8 ml min −1 as carrier solution and a 20 μl injected volume.Organic anions were detected at 215 nm, using a Varian 486 photo-diode array detector.Individual standard solutions of acetic, oxalic, citric, malic, fumaric and succinic acid, all from Sigma (Barcelona, Spain), were also used.
At the end of the experiments, the density of Bacillus spp in each rhizospheric medium was determined by dilution plating.Bacillus spp were isolated on nutrient-agar medium after heating the suspension at 80 °C for 10 min according to Tuitert et al. (1998).

Statistical analysis
The effect of microorganisms on the different measured variables was assessed by analysis of variance (ANOVA), using the General Linear Model (GLM) in the software Statgraphics Plus 5.1 (StatPoint 2000).Differences between means were assessed with Tukey's test at p < 0.05.Significant interactions between factors precluded evaluation of main effects in a combined analysis and hence comparing means between treatments for each main factor.When interactions were significant, mean comparison for the combined effect of factors was performed as recommended by Borrero et al. (2012) andde Santiago et al. (2013).Normality and homoscedasticity were always previously ascertained by using the Smirnoff-Kolmogorov and Levene test, respectively (Acutis et al. 2012), and data transformed when either or both tests were not passed.Regressions were also done with Statgraphics Plus 5.1.

Results
As expected, application of the most soluble P form (KH 2 PO4) increased P concentration in shoots and roots, total P in shoots and roots and P uptake (Tables 1 and 2).Enriching the growing medium with ferrihydrite resulted in markedly decreased root growth, total P content in roots and shoots, and P uptake.In contrast ferrihydrite resulted in significantly increased P concentration in roots and shoots, and accumulation of organic anions (OA) in the medium.Significant interactions between ferrihydrite enrichment and P source were observed for total P content in shoots, P uptake, and total Fe and total dry matter in shoots (Table 1).The effect on total P and Fe in shoots was only significant with KH 2 PO4, whereas that on DM yield was significant with both this P source and PR at 100 mg kg -1 (Table 3).Overall, the most negative consequences of enriching the growing media with Fe were observed with P applied in its most soluble form.None of the factors studied affected hydrolytic enzyme activity or CFU in the rhizosphere at the end of the experiment (Table 1).Inoculation with B. subtilis significantly increased DM yield in roots (Tables 1 and 2).The microorganism significantly increased total P in shoots and roots, and also total P uptake (Tables 1 and 2); in fact, P uptake was increased by 40 % on average relative to the control without inoculation (Table 2).By contrast, the effects of B. subtilis on P concentration in plants were not significant (Table 1).= estimated as the total P in aerial parts and roots minus P present in seeds.Only the effects of factors were considered when the factor was not involved in a significant interaction.Means followed by a different lowercase letter in a column were significantly different according to Tukey's test p < 0.05) for each factor.
Bacillus subtilis decreased organic acid accumulation in rhizospheric soil by 43% relative to non-inoculated plants (Table 2).Inoculation, however, had no significant effect on phosphatase hydrolytic activity in the rhizosphere.
The effect of the microorganism on Fe concentration in shoots was found to depend on P source as revealed by the significant interaction between both factors (Table 1); this was especially so with P applied as KH 2 PO 4 (Table 4).
The effect of the microorganism on soil pH in the rhizosphere was also dependent of P source (Table 1), and significant with KH 2 PO 4 and with PR at 100 mg kg -1 (Table 4).
Phosphorus uptake was related to total Fe in plants (data normalized by logarithmic transformation; R 2 = 0.33; p < 0.001).However, the relationship was highly significant for inoculated plants, but very weak for non-inoculated plants (Fig. 1).Phosphorus uptake and total Fe in plants were related to DM in roots (R 2 = 0.52 and R 2 = 0.43, respectively; both p < 0.001); there was, however, no significant differences in the regressions between inoculated and non-inoculated plants (results not shown).

Discussion
Phosphorus availability to plants was found to be restricted to a similar extent by a poorly soluble P source and by the presence of sorbent surfaces (Fe oxides) in the media when P was applied as soluble phosphate (Table 3).Despite the decreased amount of Fe oxides present in the growing media relative to natural soils, unsaturation of sorbent sites on Fe oxides considerably reduced the efficiency of applied P.This finding confirms that solubilization of insoluble P forms or blocking of P applied to sorbent surfaces has crucial effects on P uptake by plants and hence on fertilizer efficiency.Based on the results, P from poorly soluble forms, which are gradually released through mobilizing mechanisms acting in the rhizosphere, is less prone to irreversible adsorption onto sorbent surfaces than are soluble P forms.
Inoculation with B. subtilis QST713 proved efficient towards mobilizing P; thus, total P in shoots and roots, and P uptake, were increased relative to the control without inoculation irrespective of P source and the presence of sorbent surfaces.This is consistent with the well-known P-mobilizing abilities of other strains of this microorganism (Roi et al. 2004, Orhan et al. 2006, Mena-Violante and Olalde-Portugal 2007).Based on our results, mobilization seemingly resulted at least partially from its acidifying effect; thus, the microorganism lowered the pH slightly in two of the three P treatments.However, the increased P uptake with B. subtilis QST713 in inoculated media enriched with ferrihydrite suggests that the potential benefits of the microorganism for P use by may be in part due to blocking of sorption sites on Fe oxides through release of organic compounds forming stable complexes such as siderophores with Fe; in fact, other strains of this microorganism are known to release such compounds as a Fe-mobilizing strategy (Sansinenea and Ortiz 2011, Zhao et al. 2011, Walia et al. 2014).These organic compounds compete for sorption sites with P, thus decreasing P adsorption and increasing P absorption by plants.The hypothesis of a siderophore-mediated increased P availability is supported by the evidence of increased Fe accumulation in shoots with B. subtilis QST713 when increased amounts of P were adsorbed on Fe oxides, i.e. with KH 2 PO 4 (Table 4).Microbial siderophores are known to increase Fe uptake by plants (de Santiago et al. 2013).Also, P adsorption onto Fe oxides is known to decrease Fe uptake by plants by preventing adsorption of Fe-complexing organic compounds and hindering P dissolution as a result (Sánchez-Rodríguez et al. 2013).However, the increased Fe concentration in shoots in inoculated plants when P source was KH 2 PO 4 and the increasing trend in total Fe in shoots (p = 0.0506) suggest that microbial siderophores are more efficient in competing for sorption sites on Fe oxides with P than Fe-complexing compounds released by plants.On the other hand, the decreased concentration of low-molecular weight organic acids in inoculated media suggests that the nature of these microbial  Fe-mobilizing compounds is different from the OA excreted by roots.The compounds, usually referred to as "siderophores", are especially efficient and specific in complexing Fe relative to low-molecular organic acids excreted by plants (de Santiago et al. 2009(de Santiago et al. , 2013)).
Bacillus subtilis QST713 was effective in improving growth of cucumber plants measured as total DM .The increased growth may have been the consequence of enhanced plant nutrition.The assumption that P availability influences DM yield is supported by the fact that Fe oxides restricted the uptake by plants of applied P. P adsorption on Fe oxides implies that part of adsorbed P remains unavailable to plants (Delgado and Scalenghe 2008).Thus, P adsorption on Fe oxides results in decreased P concentration in solution and uptake by plants in Fe-enriched media.This explaines the lower DM yield in Fe-enriched media when compare with media without Fe oxides.However, increasing root growth is a common plant strategy for improving P uptake in P-deficient media (Sun et al. 2014, Wang et al. 2014); also, it is consistent with the positive relationship between root DM and P uptake by plants in observed this work.The increase in root DM resulting from the presence of B. subtilis was independent on any factors affecting P availability in the growing medium (viz., P source or ferrihydrite).Therefore, B. subtilis increased plant growth -more markedly in roots (43 %) than in shoots (29 %)-in media with a high or low P availability.This suggests that B. subtilis plays a direct role in increasing DM yield that is not necessarily mediated by improved P availability to plants.Thus, the enhanced root growth with B. subtilis QST713 increases P uptake by plants irrespective of the P availability level in the growing medium.This assumption is consistent with previous evidence of improved growth in plants inoculated with other B. subtilis strains resulting from their production of auxins (Roi et al. 2004, Lim andKim 2009).The results are also consistent with those of Orhan et al. (2006) and Turan et al. (2007), who found other Bacillus strains to increase P uptake through improved root growth.
Organic anions exudation is one response of microorganisms and plants to P scarcity in the growing medium (Jones 1998, Walker et al. 2003).Also, exudation of organic anions by cucumber roots is known to be a mobilizing strategy triggered under nutrient deficiency conditions (Kamilova et al. 2006).This may account for the increased organic anion concentrations observed here when P availability to plants was decreased by application of ferrihydrite (Table 2).The fact that oxalate was the only organic anion detetected in most of the growing media (results not shown) is consistent with previous works showing that this compound is the main anion released by plant roots under P-deficient conditions (Jones 1998, Ström et al. 2005).Organic anions released by plants or bacteria can compete for sorption sites and facilitate P desorption as a result; also they can enhance dissolution of sparingly soluble phosphates (Ström et al. 2002(Ström et al. , 2005)).However, inoculation with B. subtilis decreased the concentration of organic acids in the rhizospheric medium.This can be ascribed to consumption of this C source by B. subtilis QST713, which is consistent with previous evidence that inoculation with microorganisms can decrease the concentration of OA in the rhizosphere (García-López et al. 2015).Therefore, the benefits of B. subtilis QST713 on P nutrition of plants cannot be ascribed to the production of low-molecular weight organic acids; rather, the acids exuded by plant roots can indirectly improve P uptake by plants by acting as C sources for the microorganism.
Low-molecular organic acids play a prominent role in an Fe acquisition strategy for dicotyledonous plants.Thus, the decreased organic anions concentrations observed with QST713 provide additional evidence for its contribution to Fe uptake by plants via other, more effective Fe-mobilizing mechanisms such as the above-described release of siderophores.A significant relationship between P uptake and total Fe in plants was observed with B. subtilis but not in non-inoculated plants (Fig. 1).This likely reveal that different mechanisms affecting Fe and P availability to plants may be involved when plants are inoculated with QST713.Thus, P uptake in inoculated plants was more significantly related to Fe uptake than it was in non-inoculated plants (Fig. 1).This result suggests that P mobilization in inoculated media is more markedly dependent on Fe-mobilizing mechanisms than it is in non-inoculated media.

Conclusions
Bacillus subtilis QST713 was effective in increasing P uptake by plants irrespective of the solubility of the P source applied or the adsorption capacity of the growing medium.This benefit on P uptake can be ascribed to an increased P mobilization from growing media and to an enhanced root growth.The increased Fe accumulation in plants grown in ferrihydrite-enriched media reveals a link between Fe and P acquisition strategies, and also that P uptake is improved by adsorption of siderophores onto Fe oxides.

Fig. 1 .
Fig. 1.Log-normalized relationship between P uptake by plants and total Fe in plants; solid line for plants inoculated with Bacillus subtilis QST713 and dotted line for noninoculated plants

Table 1 .
Significance (p values) in the analysis of variance of the different variables studied in cucumber plants (Cucumis sativus

Table 2 .
Effect of the main factors on plant development, concentration and total content of P and Fe and organic acid in growing media.

Table 3 .
Effect of the interaction between P source and Fe rate on various properties of the plants and growing media PR = phosphate rock; ‡ = estimated as the total P in aerial parts and roots minus P present in seeds Means followed by a different lowercase letter in a column were significantly different according to Tukey's test at a p < 0.05 (for combinations of treatments)

Table 4 .
Effect of the significant interaction between P source and inoculation with Bacillus subtilis on Fe concentration in shoots and pH in growing media Mean, n = 10; PR = phosphate rock; Means followed by a different lowercase letter in a column were significantly different according to Tukey's test at p < 0.05 (for combinations of treatments)