The screening for ACC deaminase activity in the strains was conducted following the approach described by Glick et al. (1995), with minor modifications. The study involved the analysis of 213 diazotrophic strains isolated from various Brachiaria genotypes as part of the Embrapa project (number 02.13.08.004.00.00). The isolation, the taxonomic and partial functional characterization of these 213 strains were previously reported by Ribeiro et al. (2020).

The strains were cultivated in 5 mL of DYGS medium (Rodrigues Neto et al., 1986) and incubated at 30°C, 180 rpm, for 24 or 48 h, depending on the bacterial growth rate. Following incubation, 100 µL of the culture was transferred to new tubes containing 5 mL of LGI or NFb medium (Baldani et al., 2014) supplemented with 1 g L⁻¹ of (NH₄)₂SO₄ as a nitrogen source and incubated under the same conditions. Afterward, 100 µL of the second-round culture was transferred to fresh tubes containing 5 mL of LGI or NFB medium, but without the nitrogen source. The culture medium was then supplemented with 3 mmol L⁻¹ ACC and incubated under the previously described conditions.

Petri dishes containing Noble Agar (low nitrogen content) were supplemented with 3 mmol L⁻¹ ACC from a filtered sterilized stock solution (0.5 mol L⁻¹), which was evenly spread over the surface of the culture medium. Cultures were inoculated using a sterile cotton swab and incubated at 30°C for 48 or 60 h, depending on growth conditions. For the negative control, cultures were plated on LGI or NFb media without the addition of inorganic nitrogen or ACC substrate. A diazotrophic Herbaspirillum frisingense strain GSF30^T, recognized for its ACC deaminase activity, was used as a positive control (Rothballer et al., 2008).

The activity of ACC deaminase was assessed using the method described by Penrose and Glick (2003), which quantifies the α-ketobutyrate produced through the cleaved by the enzyme ACC deaminase. The quantification process involved measuring the absorbance of bacterial sample at 540 nm and comparing the results to a standard α-ketobutyrate curve ranging from 10 to 1000 µmol. To determine the specific activity of the cultures, protein concentration was measured using the Bradford method (Bradford, 1976).

The pre-selected strains (qualitative assays) were cultured overnight in 5 mL of DYGS medium at 30°C, 180 rpm, for 24 h. After incubation, the cells were harvested by centrifugation at 5,000 xg for 10 min at 4°C, followed by washing with NFb or LGI medium (without a nitrogen source). The bacterial pellet was then resuspended in 5 mL of NFb or LGI medium supplemented with 3 mmol L⁻¹ ACC as the sole nitrogen source. The culture was incubated for 24 h with shaking at 180 rpm at 30°C. Subsequently, the bacterial cells were harvested again by centrifugation at 5,000 xg, 4°C, for 10 min. The cells were washed twice with 5 mL of 0.1 mol L⁻¹ Tris-HCl buffer (pH 7.6). Finally, the cell suspension was transferred to a microcentrifuge tube and centrifuged at 10,000 g for 1 min.

All the supernatant was carefully removed, and the cell pellet was utilized for the enzymatic assay. The pellets were resuspended in 400 μL of 0.1 mol L⁻¹ Tris-HCl buffer (pH 8.0), followed by the addition of 20 μL of toluene and vortexing for 30 s. Subsequently, 50 μL of the toluene-treated cells were incubated with 5 μL of 0.5 M ACC at 30°C for 30 min. After incubation, 500 µL of 0.56 M HCl was added, and the mixture was vortexed and centrifuged at 10,000 xg for 5 min at room temperature. The resulting supernatant (500 μL) was vortexed with 400 μL of 0.56 M HCl and 150 μL of 2,4-dinitrophenylhydrazine reagent (0.2% 2,4-dinitrophenylhydrazine in 2 M HCl). The mixture was incubated at 30°C for 30 min, followed by the addition of 1 mL of 2 M NaOH and thorough mixing. Absorbance at 540 nm was then measured using a spectrophotometer. The cell suspension without ACC served as the negative control. Specific activity of the cultures was determined by protein quantification following the Bradford method (Bradford, 1976). ACC deaminase activity was expressed as μmol of α-ketobutyrate per mg of protein per hour.

The positive strains in the qualitative screening and those that showed results in the quantification of ACC deaminase enzyme activity were used in the detection of the acdS sequence. Genomic DNA was extracted using the commercial Wizard^® Genomic DNA Purification Kit (Promega, Madison, USA) following the manufacturer’s instructions. The concentration of genomic DNA was evaluated using a Nanodrop^® 3300 spectrophotometer (Thermo Fisher Scientific Inc., Waltham, USA). PCR reactions were performed using two pairs of primers described by Li et al. (2015): acdSf3 (5′ – ATCGGCGGCATCCAGWSNAAYCANAC – 3′), acdSr3 (5′ – GTGCATCGACTTGCCCTCRTANACNGGRT – 3′), and acdSr4 (5′ – GGCACGCCGCCCARRTGNRCRTA – 3′). Each amplification reaction was conducted in a final volume of 25 μL, consisting of 20 ng µL⁻¹ genomic DNA, 1× Taq DNA polymerase buffer (1 mM Tris-HCl, pH 9.0, and 5 mM KCl), 0.5 mM of each dNTP, 3 mM MgCl₂, 0.4 μM of each primer, and 0.1 U µL⁻¹ Taq DNA polymerase (Promega, Madison, USA). Negative control samples were prepared by replacing bacterial DNA with ultrapure water. Amplification reactions were carried out in a SureCycler 8800 thermocycler (Agilent Technologies, Santa Clara, USA) programmed for initial denaturation at 94°C for 4 min; 35 cycles at 94°C for 45 s, 53°C for 45 s, and 72°C for 1 min; followed by a final extension at 72°C for 10 min. After amplification, 2 μL of PCR product was analyzed through electrophoresis on a 1.5% (w/v) agarose gel at 90 volts (~5 V/cm) for 1 h and 30 min in 1× TAE buffer (40 mM Tris-acetate, pH 8.0, and 1 mM EDTA, pH 8.0). The gel was stained with ethidium bromide solution (0.5 μg mL⁻¹) and visualized under ultraviolet light using a KODAK^® Gel Logic Cabinet 100 photoceller (Eastman Kodak Company, Rochester, USA).

A preliminary experiment was conducted to determine the optimal PEG 8000 concentration for in vitro studies. The application of 20% PEG 8000 was highly detrimental, leading to the death of nearly all plants (data not shown). Based on these results, a subsequent experiment was performed using a reduced concentration of 10% PEG 8000.

The gnotobiotic inoculation experiment was conducted using disinfested seeds of Brachiaria ruziziensis, a genotype with low tolerance to water deficit stress. The seeds were peeled and sterilized by washing in 70% (v/v) ethanol for 3 min, followed by immersion in sodium hypochlorite (4–6% v/v free chlorine) with agitation for 10 min. After that, the seeds were rinsed three times with sterile distilled water and placed in Petri dishes containing an agar/water medium (0.5% agar supplemented with 500 mg L⁻¹ of yeast extract). The plates were initially incubated in the dark at 30°C for 24 h. Subsequently, they were transferred to a BOD incubator (model LB41, LABTEC, Londrina, Paraná, Brazil) and maintained at 30°C with a 12-h photoperiod for 4 days to ensure complete germination.

Meanwhile, the bacterial strains were inoculated in 50 mL of liquid DYGS medium and incubated under agitation at 180 rpm and 30°C, for 24 or 48 h, depending on the bacterial growth rate. Bacterial growth was quantified using the micro drop technique (Romeiro, 2007). The inoculum concentrations obtained were 10⁴ CFU mL⁻¹ for strain BR11790 and 10⁵ CFU mL⁻¹ for the other target diazotrophic bacteria. Non-contaminated seedlings were carefully removed from the agar/water medium and immersed for 1 h in the bacterial culture suspension of their respective strains: NRB032 (Stenotrophomonas maltophilia), NRB039 (Nitrospirillum amazonense), NRB058 (Pseudomonas cremoricolorata), NRB096 (Bacillus safensis), NRB123 (N. amazonense), NRB124 (Paraburkholderia silvatlantica), NRB127 (Herbaspirillum seropedicae), NRB138 (Gluconacetobacter diazotrophicus), NRB142 (P. silvatlantica), NRB223 (Azospirillum melinis) and BR11790 (H. frisingense GSF30^T). These strains were employed as they showed positive results in the qualitative screening and demonstrated activity in the quantification of the ACC deaminase enzyme. A Herbaspirillum frisingense strain GSF30^T was used as a positive control. The seedlings assigned to the control treatment were immersed in flasks containing the same volume of DYGS liquid medium for the same duration. Afterward, the inoculated seedlings were transferred to glass tubes containing 25 mL of MS medium (Murashige and Skoog, 1962) supplemented with 30 g L⁻¹ of sucrose, with the pH adjusted to 5.8. The medium was prepared both with and without 10% PEG 8000. The plants were then placed in a growth room and maintained for 30 days under a photoperiod of 16 h of light and 8 h of darkness, at a constant temperature of 25°C.

One experiment was conducted in a completely randomized design with four replications. The factors included stress induction mediated by PEG 8000 (present or absent), two seed treatments (inoculation with 11 diazotrophic strains exhibiting ACC deaminase activity and a control), and one Brachiaria genotype (B. ruziziensis). Each experimental unit consisted of a 100 mL glass tube containing 25 mL of MS medium. Analyses were performed 30 days post-inoculation by measuring plant height and the fresh biomass accumulation of leaves and roots. To compare treatment means, the Scott-Knott test was applied at a significant level of 0.05. All statistical analyses were conducted using the software ‘Sisvar’ version 5.3 (Ferreira, 2011).

To assess the Brachiaria plant colonization, a red-fluorescent derivative of NRB142 (P. silvatlantica) was constructed via transformation with plasmid pLMB426 applying the electroporation method. The transformed strain, designated NRB142 (mCherry), was cultivated in liquid or solid DYGS medium supplemented with gentamycin (80 μg mL⁻¹). Plants of B. brizantha cv. Paiaguás and B. ruziziensis with 5 days after germination were inoculated with NRB142 (mCherry). This strain was selected because it exhibited the highest performance in ACC deaminase activity quantification and showed beneficial effects in the in vitro test with PEG 8000. The gnotobiotic inoculation experiment utilized disinfected seeds of B. brizantha cv. Paiaguás and B. ruziziensis, as described in the previous section.

After germination, microorganism-free plants were removed from agar/water medium plates and transferred to glass tubes containing 25 mL of MS medium (Murashige and Skoog, 1962) supplemented with 30 g L⁻¹ sucrose and adjusted to pH 5.8 for rooting. Plants were maintained in MS medium for 30 days to promote root formation before being transferred to flasks containing 25 mL of Hoagland’s solution. Prior to transfer, inoculation with NRB142 (mCherry) was conducted in tubes designed for the inoculated treatment, using bacterial suspensions prepared at a concentration of 10⁵ CFU mL⁻¹. Bacterial growth was quantified using a Neubauer chamber.

Control treatment tubes were inoculated with PBS buffer in volumes equal to the bacterial solution. The experiment followed a completely randomized design with four replications, considering two experimental factors: inoculation with or without NRB142 (mCherry) and two Brachiaria cultivars (B. brizantha cv. Paiaguás and B. ruziziensis). Each experimental unit consisted of a 100 mL glass tube containing 25 mL of MS medium (Murashige and Skoog, 1962).

Plants were maintained in a growth chamber for 30 days under a controlled photoperiod of 16 h of light and 8 h of darkness at 25°C. Harvests were performed at 3, 7, and 14 days after inoculation (dai). Endophytic and rhizospheric bacterial populations were quantified using the micro-drop technique (Romeiro, 2007). Confocal microscopy images were obtained using the LSM 700 microscope, AxioObserver (Carl Zeiss, Jena, Germany), and processed with Zen 2.3 software (Carl Zeiss, Jena, Germany).

The methodology adapted from Glick et al. (1995) was initially used to assess the presence of ACC deaminase activity in bacterial strains. Some strains exhibited growth in LGI or NFb agar plates with ACC as the sole nitrogen source, indicating positive ACC deaminase activity. The results indicated that bacterial growth relied on ACC as its sole nitrogen source, consistent with the methodology described by Glick et al. (1995). Screening of the 213 bacterial strains isolated from Brachiaria genotypes revealed that approximately 15% possessed ACC deaminase activity. Among these, 25% were isolated from rhizospheric soil, 25% from disinfected roots, and 50% from non-disinfected roots ( Table 1 ).

The activity of the enzyme ACC deaminase was determined by quantifying the α-ketobutyrate produced during the deamination of ACC by the enzyme. In this work, 32 strains that showed growth capacity in ACC culture medium as the sole nitrogen source were selected to quantify the activity of the ACC deaminase enzyme. The results indicate that 17 out of the 32 strains exhibited ACC deaminase activity in vitro ( Table 1 ), while the remaining 15 strains tested negative for ACC deaminase activity. These findings suggest that growth in a medium with ACC as the sole nitrogen source is not sufficient to confirm that a bacterial strain possesses ACC deaminase activity. Therefore, it is essential to quantify ACC enzyme activity to verify its presence.

As expected, the positive control, H. frisingense GSF30^T, exhibited ACC deaminase activity of 9.28 μmol α-ketobutyrate mg^-1 protein h^-1 in the present assay. Another species of the same genus, H. seropedicae (NRB127), showed higher activity than H. frisingense GSF30^T, with a value of 44.83 μmol α-ketobutyrate mg^-1 protein h^-1. The species P. silvatlantica (NRB142) and G. diazotrophicus (NRB138) showed the highest ACC deaminase activities in vitro, with values of 102.52 and 89.49 μmol α-ketobutyrate mg^-1 protein h^-1, respectively. Among the results obtained, a group of bacteria presented intermediate ACC deaminase activity values, ranging from 49.0 to 16.0 μmol α-ketobutyrate mg^-1 protein h^-1. For instance, strain NRB058 (P. cremoricolorata) exhibited an activity of 49.30 μmol α-ketobutyrate mg^-1 protein h^-1 whereas the strain NRB087 (A. oryzae) showed a lower value of 16.44 μmol α-ketobutyrate mg^-1 protein h^-1. Regarding the lowest enzymatic activity values, results ranged from 8.90 to 1.98 μmol α-ketobutyrate mg⁻¹ protein h⁻¹. The lowest observed enzymatic activity was recorded for strain NRB086, with a value of 1.98 μmol α-ketobutyrate mg⁻¹ protein h⁻¹.

The predicted amplified PCR products (~ 680 bp with acdSf3/acdSr3 or ~ 760 bp with acdSf3/acdSr4) were successfully obtained for 9 bacterial strains exhibiting ACC deaminase: NRB032 (S. maltophilia), NRB058 (P. cremoricolorata), NRB086 (A. lipoferum), NRB087 (A. oryzae), NRB096 (B. safensis), NRB127 (H. seropedicae), NRB138 (G. diazotrophicus), NRB142 (P. silvatlantica) and NRB223 (A. melinis). Additionally, amplification was observed for the positive control H. frisigense GSF30^T (BR11790). In contrast, no amplification of the acdS gene was detected in the negative control (blank sample). An agarose gel electrophoresis illustrating the respective amplified product is shown in Figure 1 .

The growth of B. ruziziensis in the presence of 10% PEG8000 was considerably decreased compared to control treatment (no PEG8000), which exhibited higher values for the analyzed variables ( Table 2 ). There was a statistically significant difference (p <0.05) was observed in plant size between the control treatment and those subjected to water deficit stress. However, under water deficit stress, inoculation with strains NRB223 (A. melinis), BR11790 (H. frisigense GSF30^T), NRB142 (P. silvatlantica), NRB032 (S. maltophilia) and NRB127 (H. seropedicae) lead to an increase in plant size, statistically differing from the other strains and the uninoculated plants. Among these plants inoculated with strain NRB223 (A. melinis) exhibited the greatest height (45.00 cm), while the shortest height (9.00 cm) was recorded in the uninoculated plants.

In the absence of PEG8000, no statistically significant differences were observed among the inoculated treatments and the control, except for strains NRB123 (N. amazonense) and NRB127 (H. seropedicae), which showed smaller plant sizes ( Table 2 ). Plants inoculated with strains NRB223 (A. melinis), NRB124 (P. silvatlantica), NRB032 (S. maltophilia) showed a higher increase in plant size, reaching up to 59.40 cm. In contrast, plants inoculated with strains NRB127 (H. seropedicae) and NRB123 (N. amazonense) displayed comparatively smaller sizes, measuring 46.70 and 42.60 cm, respectively.

A statistically significant difference (p <0.05) was observed in fresh biomass accumulation both under control (without PEG8000) and in treatments subjected to water deficit stress (with PEG 8000). Under control conditions, plants inoculated with strain NRB039 (N. amazonense) showed the highest fresh biomass accumulation (6.20 g), differing statistically from the other inoculated strains. In contrast, plants inoculated with the strain NRB127 (H. seropedicae) accumulated the lower fresh biomass accumulation (1.70 g). Under water deficit stress, plants inoculated with strain NRB223 (A. melinis) accumulated the highest fresh biomass (1.50 g), followed by those inoculated with strains NRB142 (P. silvatlantica), BR11790 (H. frisigense GSF30^T), and NRB124 (P. silvatlantica). These inoculated plants differed statistically from plants inoculated with other strains and the uninoculated plants, which accumulated only 0.03g of biomass.

The ability of strain NRB142 (P. silvatlantica) to colonize seedlings of B. brizantha cv. Paiaguás and B. ruziziensis was investigated under in vitro conditions. The NRB142 strain was selected for this experiment as it demonstrated the highest activity in the quantification of ACC deaminase enzyme activity. Additionally, this strain also contributed to plant tolerance against water deficit stress, induced by PEG in vitro. The strain was successfully labeled with the plasmid harboring the mCherry gene, which remained stable throughout the colonization study, as confirmed by bacterial counting and microscopy analysis conducted at 3, 7, and 14 days after inoculation (dai).

Bacterial counts showed that non-disinfected roots (NDR) presented a higher bacterial colonizing the root system compared to disinfested roots (DR) in both Brachiaria genotypes ( Table 3 ). Differences in bacterial colonization were observed between both genotypes. At 3 dai, the Paiaguás genotype showed a greater bacterial population in both NDR or DR root system. By day 7, this pattern persisted for NDR, while that disinfected roots (DR) of the Ruziziensis genotype showed a higher bacterial population (4 × 10⁴ cells g^-1 fresh tissues). By 14 dai, a highest colony forming unit (CFU) g^-1 fresh tissue was observed for NDR of the Ruziziensis genotype, whereas DR of the Paiaguás genotype exhibited the highest bacterial count. Despite these differences, results confirmed that strain NRB142 (P. silvatlantica) effectively colonizes both Brachiaria genotypes in numbers relatively high, including the root interior. Despite these differences, results confirm that strain NRB142 effectively colonizes both Brachiaria genotypes, including root interior colonization.

Microscopy analyses further validated these findings, showing significant bacterial aggregations attached to Brachiaria roots. Red fluorescent NRB142 (mCherry) cells were observed colonizing roots of B. brizantha cv. Paiaguás and B. ruziziensis ( Figures 2B, D, F, H, J, L ). In contrast, no fluorescent bacteria were detected in non-inoculated control plants ( Figures 2A, C, E, G, I, K ). The bacterial counting (CFU) analysis corroborated these observations, as no bacterial colonies developed on plates containing DYGS culture medium inoculated with dilutions from macerated roots of control plants.