The ability of the strain Bradyrhizobium elkanii SEMIA 587 to use phytate was tested at pH 6 and 7 by inoculating the following Angle medium (Angle et al., 1991) modified by Richardson and Hadobas (1997): 1 mM KNO₃, 2 mM MgSO₄7H₂O, 4 mM CaSO₄, 55 mM glucose, 50 mg L⁻¹ thiamine-HCl, 0.5 mL L⁻¹ Fe-citrate 1%, 0.2 mL L⁻¹ micronutrient solution containing per liter (2.82 g H₃BO₃, 98 mg CuSO₄5H₂O, 3.08 g MnSO₄H₂O, 0.29 g NaMoO₄2H₂O, 4.41 g ZnSO₄7H₂O) and 6 mM Na-phytate (Sigma, ref P0109). The optical density at 600 nm was measured every 24 h during seven days (168 h). The above-mentioned medium supplemented with 1.5 mM KH₂PO₄, instead of Na-phytate, was used as control. Three replicates for both growing conditions were used.

The design of the rhizobial HAP phytase primers was performed online at the National Center of Biotechnology Information (NCBI, http://blast.ncbi.nlm.nih.gov/Blast.cgi). The highest hit by BLASTp to rhizobiaceae group of a known amino acid sequence of the Escherichia coli HAP phytase (GenBank accession number: AAN28334.1) was obtained with a conserved hypothetical HAP-type phytase of Azorhizobium caulinodans (GenBank accession number: 5692057 AZC_4319). The gene sequence of this hypothetical HAP-type phytase was obtained from A. caulinodans full genome (GenBank accession number: AP009384.1). Thereafter, the following pair of primers were designed with the primer-BLAST tool from this HAP gene-sequence: forward primer (HAPfor), 5′CAGTTCACGCCAAAGATGCC3′; reverse primer (HAPrev), 5′CGCGTATGGTCCATCCTGAA3′. Reverse primer containing three consecutive mismatched bases near its 3′ terminus can hybridize to the target RNA at a temperature (in this study: 42°C) which is not high enough for DNA denaturation. Furthermore, due to the lower affinity between this mismatched primer and genomic DNA (gDNA), it is a very inefficient primer for gDNA amplification at high annealing temperatures (in this study: 62°C). Consequently, the mismatched primer can be used to selectively amplify the newly synthesized complementary DNA (cDNA), even in the presence of high quantities of gDNA (Koo and Jaykus, 2000).

HAP designed primer pair along with a primer pair that amplifies BBP described by Farhat et al. (2008) (BPPfor, 5′GATGCAGCTGATGATCCTGCG3′; BPPrev, 5′ATTTTCTCCGTCCTGTGCGAC3′), were used to amplified B. elkanii SEMIA 587 phytases expressed in presence of phytate as unique P source. B. elkanii SEMIA 587 RNA and gDNA were extracted with the RNeasy Plus and DNeasy Plant Mini Kits (QIAGEN), respectively. RNA was extracted from B. elkanii SEMIA 587 cultivated in Angle medium supplemented with Na-phytate and KH₂PO₄.

Retrotranscription reactions were carried out in 25 μL, but first, 1 μg RNA and 3 μL of 10 μM HAP or BPP reverse primers were incubated at 70°C during 5 min and at 4°C for 2 min. Then, 5 X Buffer (Promega), 10 mM of each dNTP, 0.1% bovine serum albumin (BSA), and 200 U Moloney murine leukemia virus (M-MLV) reverse transcriptase H- (Promega), were added and incubated at 42°C during 60 min.

A touchdown PCR was performed with HAP and BPP primers and with cDNA and gDNA as templates, to test if at different temperatures it was able to amplify a phytase. PCR reactions were carried out in 25 μL volumes containing 1 μL target, 10 X Buffer (Invitrogen), 50 mM MgCl₂, 10 mM of each dNTP, 10 μM forward primer (HAP or BPP), 20 μM reverse primer (HAP or BPP), 0.1% bovine serum albumin (BSA), 5 U Taq Polymerase (Invitrogen), under the following thermal conditions: 95°C for 2 min, 30 cycles of 95°C for 30 s, 70°C for 45 s, and gradually reduced 1°C till reaching 60°C, 72°C for 45 s, with a final extension at 72°C for 3 min.

When PCR products were obtained in the touchdown PCR, a gradient PCR was performed with cDNA and gDNA as templates, to obtain the optimal hybridization temperature of primers. PCR reactions were carried out in 25 μL volumes containing 1 μL target, 10 X Buffer (Invitrogen), 50 mM MgCl₂, 10 mM of each dNTP, 10 μM forward primer, 20 μM reverse primer, 0.1% BSA, 5 U Taq Polymerase (Invitrogen), under the following thermal conditions: 94°C for 3 min, 35 cycles of 94°C for 30 s, 64-59°C for 45 s, 72°C for 45 s, with a final extension at 72°C for 3 min. Consequently, the optimal hybridization temperature of primers HAP was 62°C.

All amplification products were checked by electrophoresis in 1% agarose gel.

Two commercial soybean cultivars, namely N5909 (Nidera) and SR532 (Santa Rosa), were used in this study. SR532 and N5909 have a short and long production cycle, respectively. B. elkanii SEMIA 587 was used to inoculate these two cultivars. Before sowing, the soybean seeds were surface sterilized, first with 95% ethanol for 3 min and afterwards, with 3% NaOCl for 3 min more. Then, seeds were rinsed five times with sterile distilled water. Later, seeds were imbibed in sterile distilled water for 4 h, and then transferred for germination on sterile 15 g L⁻¹ agar-water plates (Beyhaut et al., 2006). After germination, inoculation was performed by 30 min soaking 10-day-old seedlings in a suspension of B. elkanii SEMIA 587 containing ~10⁸ cfu mL⁻¹. The inoculum was prepared by growing the strain SEMIA 587 in yeast extract mannitol broth during four days at 28°C at 120 rpm.

Twenty inoculated seedlings were transferred into each container, with base dimensions 0.2 × 0.4 m and height 0.2 m, refilled weekly with 40 L of Vadez et al. (1996) nutrient solution. In order to avoid initial N deficiency and ensure optimal nodulation, urea was supplied at 1 mmol plant⁻¹ during the initial 15 days of growth. Thereafter, the plants were grown in N-free solution. They were grown under both deficient and sufficient P supply in hydroaeroponic conditions with a permanent flow of 400 mL min⁻¹ of compressed air to ensure the oxygenation of the culture solution (Hernandez and Drevon, 1991). P was supplied weekly in the form of KH₂PO₄, at 75 vs. 250 μmol plant⁻¹ as P deficiency vs. P sufficiency, respectively (Vadez et al., 1996). The pH was adjusted to 6.8 with 0.2 g L⁻¹ CaCO₃ (Hernandez and Drevon, 1991). The experiment was carried out in a greenhouse, 28/20°C under 16/8 h day/night cycle, with an additional light supply of 400 μmol photons m⁻² s⁻¹ and 70% relative humidity during the day.

Forty two days after transfer (DAT), soybean plants were harvested and shoot, roots and nodules were separated. Dry weights (DW) were determined after drying at 70°C for three days. Shoot accumulated P was determined by the vanado-molybdate method in Laboratorio Oriental (http://labo.com.uy/).

The efficiency in use of the rhizobial symbiosis (EURS) was estimated by the slope of the regression model of shoot biomass as a function of nodule biomass (y = ax + b), where a corresponds to the EURS and b corresponds to the plant biomass production without nodules (Zaman-Allah et al., 2007).

The in situ RT-PCR was made according to Lazali et al. (2013) and Maougal et al. (2014b). Three nodules of 3 mm diameter of three plants of each cultivar and P treatment were carefully detached from roots at 42 DAT. They were thoroughly washed with diethyl pyrocarbonate (DEPC) treated water, then fixed in 4% paraformaldehyde, 45% ethanol and 5% acetic acid, kept for 2 h under vacuum, and stored overnight at 4°C. Fixed nodules were extensively rinsed with two baths of DEPC treated water during 5 min each, two baths of 1 X phosphate buffered saline (PBS, 5 mM Na₂HPO₄, 300 mM NaCl, pH 7.5) during 10 min each, one bath of 1 X PBS plus 0.2% glycine during 10 min and one last bath with 1 X PBS during 10 min.

Thereafter, the fixed nodules were dehydrated in baths of increasing concentrations of ethanol and butanol: one bath of 50% ethanol during 30 min; two baths of 70% ethanol during 30 min and 1 h; one bath of 70% of ethanol overnight at 4°C; two baths with 100% ethanol for 1 h; and in the end, at least, three baths with 100% butanol during one week. Then, the nodules were included in paraffin and cut into transversal 12 μm thick sections using a microtome (Micro-cut H1200 Vibrating Microtome). The resulting sections were placed on slides treated with silane to increase cell adhesion. Thereafter, the sections were deparaffinated through three baths of Safesolv (Q Path®) for 10 min each, one bath of Safesolv during 5 min; three baths of 100% ethanol during 10 min each; one bath of 70% ethanol during 5 min; one bath of 50% ethanol during 5 min. This was continued by section rehydration with one bath of 1 X PBS during 10 min, one bath of 1 X PBS plus glycine 0.2% during 2 min, one bath of 1 X PBS during 10 min and one bath of 1X PBS for 5 min at 65°C.

The first cDNA strand was synthesized from total RNA of nodular sections, that were incubated at 42°C during 1 h in 100 μL reverse transcriptase mix containing 5 X Buffer (Promega), 10 mM of each dNTP, 10 μM HAPrev, 0.1% BSA, and 200 U M-MLV reverse transcriptase H- (Promega). Negative controls (NRT) were prepared by omitting the reverse transcriptase. Afterwards, the reverse transcriptase mix was removed, by adding two times 1 mL of 1 X PBS and incubating the slides two times more in 50 mL of 1 X PBS during 10 min. PCR reactions were carried out in 100 μL volumes containing 10 X Buffer (Invitrogen), 50 mM MgCl₂, 10 mM of each dNTP, 1 mM digoxigenin-11-uridine-triphosphate (Dig-11-dUTP), 10 μM HAPfor, 20 μM HAPrev, and 5 U Taq polymerase (Invitrogen), under the following thermal conditions: 35 cycles of 94°C for 30 s, 62°C for 45 s, and 72°C for 45 s, with a final extension at 72°C for 3 min.

For detection of the amplified cDNA in the nodules, the PCR mix was removed, and samples were washed in 100 μl of blocking solution under gentle agitation in the dark at 37°C. The blocking solution was replaced by 100 μl of alkaline phosphatase conjugated anti-dioxygenin Fab fragment (Roche Diagnostics) diluted 1:1000 in 2% BSA. The samples were incubated at room temperature for 90 min, then washed three times to remove excess antibody. Detection of alkaline phosphatase was carried out using an ELF-97® endogenous phosphatase detection kit (Molecular Probes). The ELF substrate was diluted 1:40 in the alkaline detection buffer (Molecular Probes), vigorously shaken, and then filtered through a 0.22 μm filter (Millex®-GV, Millipore) to remove any aggregates of the substrate that may have formed during storage. Samples were incubated in 20 μL ELF substrate–buffer solution for 20 min in the dark and transferred to washing buffer (1 X PBS, 25 mM EDTA, and 5 mM levamisole, pH 8.0). Three washings of 1 min were performed. Samples were mounted for observations with a BX61® microscope (Olympus) equipped with an epifluorescence condenser, a Hoechst/DAPI filter set configured with an excitation filter of 360–370 nm, a dichroic mirror of 400 nm and an emission of 420 nm, and a gray View II® camera (ORCA AG). Image analysis was performed using ImageJ software (Schneider et al., 2012) as an image analysis program. The intensity of the fluorescent signal emitted by the rhizobial phytase transcript was measured as number of green pixels per image.

Shoot, nodules and root DW, shoot accumulated P and intensity of the fluorescent signal emitted by the rhizobial phytase transcript were analyzed through analysis of variance (ANOVA) and tested for differences between cultivars and P growing conditions using a post-hoc Tukey's HSD test (p < 0.05) in R3.1.3 software (https://www.r-project.org/) using the “agricolae” package (de Mendiburu, 2020). The relationship between nodule and shoot biomass was tested by regression analysis.

B. elkanii SEMIA 587 showed similar growth curves when it was cultivated in Angle medium supplemented with Na-phytate or KH₂PO₄, as P sources, at both pH evaluated (6 and 7) (Figures 1A,B). Therefore, this strain was able to mineralize phytate and to use it as a P source.

Two primers pairs, that amplify two types of phytases, HAP and BPP, were tested through a retro-transcription and a subsequent touchdown PCR in B. elkanii SEMIA 587. An amplification product was obtained when SEMIA 587 was cultivated with Na-phytate as P source and when HAP primers were used, but not when BPP primers were used (Figure 1C). When SEMIA 587 was cultivated with KH₂PO₄ as P source, no amplification products were obtained either when HAP or BPP primers were used (Figure 1C). Additionally, we demonstrated that HAP gene was not amplified from gDNA with these HAP primers (Figure 1C), due to the lower affinity between the mismatched primer and the gDNA at the used annealing temperature (62°C).

Cultivar N5909 cultivated under P sufficiency showed more shoot, nodules and root DW than under P deficiency (p = 0.0001, p = 0.0007, p = 0.0300, respectively), whereas no significant differences of shoot, nodules and root DW under P sufficiency and P deficiency in cultivar SR532 were found (Figure 2).

The EURS was estimated as the regression slope of shoot biomass as a function of nodule biomass, whenever the correlation between both parameters was significant (up to R² = 0.719). Under P sufficiency, SR532 was the most efficient, and showed similar EURS under P deficiency, whereas N5909 increased its EURS under P deficiency (Figure 3).

Cultivar N5909 accumulated more P in its shoots under P sufficiency than under P deficiency (p = 0.0008), whereas cultivar SR532 did not show any significant difference in shoot accumulated P between P treatments (Figure 4).

Using the In situ RT-PCR technique we were able to show the transcript abundance of the rhizobial phytase in nodular sections of soybean cultivars N5909 and SR532 inoculated with B. elkanii SEMIA 587 cultivated under P sufficiency and P deficiency (Figures 5A–D). The NRT nodules not subjected to reverse transcription did not display any fluorescent signal (Figures 5E,F), confirming that fluorescence was not due to methodological artifacts. The rhizobial phytase was transcripted in all bacteroidal symbiosomes in the infected zone, with a high transcript abundance rate since the fluorescent signal, as assessment of number of transcripts, was intense for all the treatments (Figures 5A–D).

In the case of cultivar SR532, the fluorescent signal was more intense under P deficiency in comparison to P sufficiency (p < 0.0001), whereas in the case of cultivar N5909, there were no significant differences between the intensity of the fluorescent signal under P deficiency or sufficiency. The intensity of the fluorescent signal in SR532 under P deficiency was equal to the intensity of this signal in N5909 under both P conditions (Figure 6). The ANOVA showed a significant interaction between the factors P supply and cultivar for the variable intensity of fluorescent signal (p = 0.0010).