The Microphenotron: a novel method for screening plant growth-promoting rhizobacteria

Isolation of rhizobacteria

Bacterial strains were isolated from the rhizospheric soil of Acacia Arabica (L.) that inhabited the saline areas of the Pothohar salt range, Pakistan. The rhizospheric soil samples were collected in clean bags. Samples were processed within 24 h by making serial dilutions in sterilized distilled water and inoculated on L-agar (Luria-Bertani) plates supplemented with four different NaCl concentrations i.e., 0.25, 0.5, 0.75 and 1 M. Pure cultures were obtained by quadrant streaking after picking distinct colonies from inoculated plates.

16S rRNA gene sequencing

The taxonomic status of bacterial strains was confirmed by 16S rRNA gene sequencing. DNA was isolated from freshly grown bacterial cultures by using the Genomic DNA Purification Kit (Promega, Madison, WI, USA) according to the manufacturer’s instructions. A fragment (1.5-kb) of the 16S rRNA gene was amplified by using forward and reverse primers (Johnson, 1994). PCR amplification was accomplished in Thermocycler Primus 96 (PeQLab, Erlangen, BY, Germany) as described previously (Raheem et al., 2018). The amplified PCR products were purified by using the QIAquick Gel Extraction Kit (Venlo, Netherlands). Samples were sequenced by using 27f and 1522r primers by sending purified PCR products to Eurofins, United Kingdom (UK). For final identification, sequences of the 16S rRNA gene were compared with already deposited sequences in the GenBank through BLAST analysis. Phylogenetic relationships among bacterial strains were determined by aligning sequences using the Neighbor-Joining method with 1,000 bootstrap replicates. The tree was plotted using MEGA 7 software.

Auxin production and bacterial growth

Auxin production ability of bacterial isolates was determined in the presence and absence of precursor L-tryptophan. Strains were grown in 20 ml L-broth medium supplemented with 0.5, 1, 1.5 and 2 M NaCl in 50 ml Erlenmeyer flasks. L-broth was also supplemented with the filter-sterilized solution of L-tryptophan to a final concentration of 500 µg ml⁻¹. All inoculated flasks were incubated at 37 °C for 72 h at 120 rev. min⁻¹. After incubation, cells were removed from stationary phase cultures by centrifugation at 2,300×g for 15 min. One ml of supernatant was mixed with 2 ml of Salkowski’s reagent. The contents in the test tube were allowed to stand in dark for half an hour for color development. The intensity of the color was measured at 535 nm by a spectrophotometer (CECIL, CE 7200).

Ultra-performance liquid chromatography

The quantification of indolic compounds in bacterial extracts was accomplished by ultra-performance liquid chromatography (UPLC). The L-broth medium was amended with 500 µg ml⁻¹ L-tryptophan and incubated with bacterial strains at 30 °C for 72 h as mentioned above. Afterward, cells were separated from broth medium by centrifugation at 2,300×g for 15 min. Cell-free supernatant (10 ml) was further processed for the extraction and quantification of indolic compounds as described earlier (Raheem et al., 2018). The UPLC system of Waters ACQUITY H-Class (Milford, Massachusetts, USA) coupled with FD-detector (λ_ex = 280 nm, λ_em = 350 nm) was used for the detection of indolic compounds. The mobile phase was 18% acetonitrile prepared in 0.1% acetic acid at a flow rate of 0.3 ml min⁻¹. Detection of indolic compounds was performed by comparing the retention time of peaks in the samples and the mixture of authentic auxins.

ACC-deaminase activity

Screening of 1-aminocyclopropane-1-carboxylate (ACC) deaminase activity was based on the method used by Li et al. (2011). Strains were cultivated in 5 ml L-broth at 37 °C for 24 h and cell cultures were harvested in microcentrifuge tubes at 8,000×g for 5 min. Cell pellets were washed with 1 ml of liquid DF medium. Afterward, cultures were re-suspended in DF medium and placed on a shaking incubator at 37 °C for 24 h. For comparison, 2 ml un-inoculated DF-ACC medium was used as control. Next, 1 ml culture was centrifuged and 100 µl supernatant was transferred to a new centrifuge tube. Then, the supernatant was diluted to 1 ml with liquid DF medium. For a 96-well PCR-plate assay, 60 µl diluted supernatant was mixed with 120 µl of ninhydrin reagent. The Chimney-top polystyrene heat resistant PCR plate (Thomas Scientific, Swedesboro, NJ, USA) was heated on a water bath for 30 min. For comparison, 10 replicates for un-inoculated controls and samples were processed. The Riemann’s Purple color of each treated and control tube indicated the ACC-deaminase activity. For optical density, 100 µl samples were transferred to the microtiter plate and the absorbance was recorded at 570 nm with the microtiter plate reader spectrophotometer (Epoch BioTek, Winooski, VT, USA).

Phosphate solubilization

Screening for phosphate solubilization was performed following the method described by Linu, Stephen & Jisha (2009). Selected strains were streaked on the Pikovskaya medium and incubated at 28 °C for 7 days.

Hydrogen cyanide

Screening for bacterial hydrogen cyanide (HCN) production was determined following Lorck (1948) and Castric (1977) with few modifications. Glycine was added to nutrient agar, sterilized by autoclaving, and poured into Petri plates. A lawn of fresh bacterial culture was made on modified agar plates. Whatman filter paper No. 1 was soaked in a solution containing 0.5% picric acid supplemented with 2% sodium carbonate. The filter paper was placed at the top of the culture plates. These plates were incubated at 30 °C for 4 days and sealed with parafilm to avoid contamination. After incubation, the development of orange to red color was recorded.

Plant materials

The wild-type Arabidopsis thaliana L. accession Col. N6000 and the mutant lines aux1, aux4, etr1, and EIN2.5 were obtained from the European Arabidopsis Stock Centre Nottingham, United Kingdom. The mutant line aux1 and aux4 were auxin resistant and EIN2.5 and etr1 were ethylene insensitive.

Screening rhizobacteria by microphenotyping system

Strips of eight PCR tubes (Frame StripTM; 4titude, Wotton, United Kingdom), supported in 96-well PCR boxes (Thermo Fisher Scientific, Waltham, MA, USA), were filled with approximately 300 µl of autoclaved B5 Medium. For the Microphenotron bioassay, protocols of Forde et al. (2013) and Burrell et al. (2017) were slightly modified to screen the effect of bacterial metabolites on plant growth. Seeds of wild-type Columbia N6000 were surface sterilized and sown onto the agar surface (6–12 seeds per tube), stratified in the dark at 3 °C for 2–4 days. Afterward, seeds were transferred to the growth room at 22 °C with a 16 h photoperiod at a light intensity of 38 µmol m⁻¹ s⁻¹. The PCR boxes were kept in propagators lined with a moistened absorbent paper towel to maintain humidity. After 2–3 days, the bottom of each tube 2 mm was excised using a guillotine, and the strips were transferred to 96–well microtitre plates containing 150 µl of bacterial suspension in B5 medium. The cell density of bacterial suspension was adjusted to 10⁷ CFU ml⁻¹. A single lane of PCR plate that contained eight tubes was used for a single strain. Similarly, a single lane of eight tubes was used for the B5 medium as control. For each strain and control, three separate strips of eight PCR tubes were kept and the experiment was repeated three times. Tubes were incubated for 2 days to allow microbial metabolites to diffuse into the root zone. For high-throughput screening, the roots were observed for 3–4 days by removing the strips individually and scoring visually for the presence or absence of the distinctive rhizobacterial elicited root phenotype. A Canon 600D camera with a 60–mm f/2.8 lens was used for lower power imaging (Canon, Woodhatch Reigate Surrey, UK). PCR strips for each strain or control were photographed in a group of four tubes i.e., in tetrads. To measure root growth rates, the position of the tip of the most advanced primary root was marked at intervals on the side of each tube, images were captured using a Canonscan 4200 flat-bed scanner and analyzed using ImageJ 1.47v software.

Bacteria-Arabidopsis pot trials

The effectiveness of the Microphenotron system was determined by performing pot trials with A. thaliana L. (wild type) under greenhouse conditions. For pot trials, an autoclaved mixture of compost, sand, and vermiculite was used in 6:1:1, respectively. Pots were filled with medium and placed in a plastic tray adjusted with 4–5 mm autoclaved water at the bottom. Seeds were sterilized with 0.1% mercuric chloride for 3 min and placed in a fridge for 2 to 3 days to break the dormancy. Three seeds per pot were sown and placed at control temperature and light intensity and duration. For each strain, three pots were placed, and the experiment was repeated three times. The temperature of the growth chamber was 24 °C and 8/16 h dark/light duration. After 3–4 days, pots were inoculated with 10 ml bacterial suspension adjusted to 10⁷ CFU-ml⁻¹ with a sterile disposable syringe. After flowering (15–19 days), plants were harvested to record rosette fresh weight, leaf area, number of rosette leaves, and plant fresh weight.

The efficacy of the micro-phenotyping system was also evaluated by performing pot trials with the mutant line of A. thaliana. The involvement of bacterial phytohormones in plant growth promotion was elucidated by using four different mutant lines aux1 (auxin resistant), axr4 (auxin resistant), ein2.5 (ethylene insensitive), and eir1 (ethylene sensitive). These mutant lines were sown in plastic pots filled with potting media as mentioned above.

Statistical analysis

Data for bacterial auxin production and plant growth parameters were subjected to statistical analysis by using SPSS 20 software. Data were subjected to analysis of variance (ANOVA). Finally, mean values were separated by using Duncan’s multiple range test (P ≤ 0.05). Correlation coefficients between bacterial auxin production and cell densities at different salinity levels were also calculated (P ≤ 0.05).

Identification of rhizobacteria

Sequencing analysis indicated that bacterial strains S-6, S-26, and S-50 showed similarities with genus Bacillus. All the Bacillus strains were clustered together in the phylogenetic tree (Fig. 1). S-10 and S-24 showed relatedness to genus Enterobacter. Strains S-7 and S-11 recorded similarity with genus Psychrobacter. Bacterial strains S-3, S-29, and S-80 were related to the genus Prolinoborus, Moraxella, and Pseudomonas, respectively. The sequences have been submitted to GeneBank under accession numbers KJ011870 to KJ011879.

Auxin production and bacterial growth

Bacterial strains recorded variable auxin production ability at different salt stresses (Figs. 2A–2D). Overall, auxin production was negatively correlated with bacterial cell densities with increasing salinity levels (r = −1**, P ≤ 0.05). Strains of B. endophyticus S-6, E. asburiae S-24, and B. thuringiensis (S-26, S-50) S-50, grew well on 1 M NaCl stress. However, increasing NaCl concentrations recorded a decrease in bacterial cell densities. The auxin production ability of these strains was also decreased with increasing NaCl levels. All strains showed a variable IAA level in the culture supernatants. For instance, E. aerogenes S-10 and B. endophyticus S-6 recorded 105 µg IAA ml⁻¹ at 1 M NaCl (Fig. 2). Similarly, strains S-11, S-24, S24 of Psy. Alimenterius, E. asburiae and B. thuringiensis also recorded significant auxin production.

Ultra-performance liquid chromatography

Ultra-performance liquid chromatography (UPLC) confirmed the production of three kinds of indolic compounds in the bacterial culture supernatants: indol-3-acetic acid (IAA), indole-3-lactic acid (ILA) and indol-3-carboxylic acid (ICA). For IAA, maximum production was shown by Psy. alimentarius S-7 (22 µg ml⁻¹) and E. asburiae S-24 (17 µg ml⁻¹). On the other hand, Psy. alimentarius S-11, E. aerogenes S-10 and E. asburiae S-24, respectively, produced 21, 13, and 10 µg ml⁻¹ ICA. Similarly, for ILA, the highest production was recorded with Psy. alimentarius S-7 (8 µg ml⁻¹), B. thuringiensis S-50 (4 µg ml⁻¹), and Pro. fasciculus S-3 (4 µg ml⁻¹). Figures 3A and 3B showed the production of different indolic compounds in bacterial culture extracts and its comparison with respective standard peaks.

Additional plant growth promoting traits

In addition to auxin production, the majority of the bacterial isolates also showed positive results for phosphate solubilization, ACC-deaminase activity, and HCN production. For example, Psy. alimentarius S-7, E. asburiae S-24, B. thuringiensis (S-26, S-50), M. pluranimalium S-29, and P. stutzeri S-80 were able to solubilize phosphate. All the strains were able to degrade ACC and were positive for HCN production. The highest ACC-deaminase activity was exhibited by B. endophyticus S-6 (241 nmol h⁻¹), E. aerogenes S-10 (197 nmol h⁻¹) and P. stutzeri S-80 (190 nmol h⁻¹) (Fig. 4).

The Microphenotron bioassay with A. thaliana

The screening bioassay through the 96-well PCR plate method with wild-type A. thaliana (Col. N6000) showed an increase in root and shoot length (Fig. 5). Strains S-26 and S-50 of B. thuringiensis recorded around 2-folds increase in root growth (Fig. 5A, Fig. S1). Similarly, M. pluranimalium S-29 and P. stutzeri S-80 also recorded significant root growth response over control. All the test strains increased root proliferation of A. thaliana in comparison with control (Fig. 6). Overall, above mentioned strains also recorded consistent results for shoot length in the Microphenotron bioassay (Fig. 5B).

Pot trials with A. thaliana L. wild-type Col. N6000

For A. thaliana wild-type Col. N6000, significant increases in leaf area, number of leaves, rosette fresh weight, and plant fresh weight were observed with bacterial inoculation as compared with un-inoculated control. A significant increase in leaf area was shown by E. asburiae S-24 (38%), M. pluranimalium S-29 (49%), and B. thuringiensis S-50 (37%) (Fig. 7A), whereas the maximum number of leaves over control were recorded by P. stutzeri S-80 (52%) (Fig. 7B). Rosette fresh biomass was observed with Psy. alimentarius S-7 and Pro. fasciculus S-3 (Fig. 8A). On the other hand, statistically significant improvements for plant fresh biomass were recorded with M. pluranimalium S-29, and Psy. alimentarius S-7, over water-treated control (Fig. 8B).

Effects of rhizobacteria on hormonal mutant lines of Arabidopsis thaliana L.

Bacterization of A. thaliana aux1 with M. pluranimalium S-29 and E. aerogenes S-10 showed 75% and 62% increase in rosette fresh biomass, over control (Table 1). For the leaf area, Psy. alimentarius S-11 (25%), P. stutzeri S-80 (24%), and B. thuringiensis S-50 (23%) showed significant improvements. Similarly, for the number of leaves, strains S-3 and S-29 of Pro. fasciculus and M. pluranimalium recorded 84%, and 46% increases, respectively. Statistically significant increases of 1.95, and 1.67 folds in plant fresh biomass were recorded with M. pluranimalium S-29, and B. thuringiensis S-50. However, Psy. alimentarius S-11, E. asburiae S-24, and B. thuringiensis S-26 did not show the statistically significant response for respective growth parameters.

B. thuringiensis (S-26, S-50), and P. stutzeri S-80 significantly enhanced rosette fresh weight in the auxin resistant mutant (Table 2). For instance, treatment with P. stutzeri S-80, and B. thuringiensis (S-26, S-50) showed maximum improvements in leaf area. Similarly, for the number of leaves, Pro. fasciculus S-3, M. pluranimalium S-29, and B. thuringiensis S-50 showed significant increases. For fresh biomass, B. thuringiensis S-50 (189%), P. stutzeri S-80 (162%), and E. aerogenes S-10 (86%) were the most promising. However, axr4 treatment with E. asburiae S-24 did not record any response for rosette fresh weight or leaf area, over control.

Inoculation of ein2.5 with M. pluranimalium S-29, P. stutzeri S-80, and B. thuringiensis S-50 showed up to one-fold increase for rosette fresh biomass, over control (Table 3). Similarly, the leaf area was enhanced with B. thuringiensis S-26, E. aerogenes S-10, and M. pluranimalium S-29. A statistically significant increase in the number of leaves was recorded with M. pluranimalium S-29 (49%), B. thuringiensis S-26 (34%), and P. stutzeri S-80 (23%). For fresh biomass, maximum improvements were recorded with B. thuringiensis S-50, P. stutzeri S-80, and Psy. alimentarius S-11, respectively when compared with un-inoculated control. However, a statistically insignificant response was recorded for rosette fresh weight with Psy. alimentarius S-7, Psy. alimentarius S-11, and E. asburiae S-24.

Bacterization of etr1 stimulated rosette growth with P. stutzeri S-80, B. thuringiensis S-50, and Psy. alimentarius S-11, over control (Table 4). For leaf area, P. stutzeri S-80, B. thuringiensis S-50, and Psy. alimentarius S-11 recorded maximum improvements. For the number of leaves, Pro. fasciculus S-3, E. aerogenes S-10 and M. pluranimalium S-29 were the most effective. In the case of plant fresh biomass, B. thuringiensis (S-26, S-50), and P. stutzeri S-80 showed up to 1.9 fold increases over respective un-inoculated controls. However, in case of rosette or plant fresh weight, a statistically insignificant response was recorded with Pro. fasciculus S-3, B. endophyticus S-6, Psy. alimentarius S-7, E. aerogenes S-10, and B. thuringiensis S-26, over untreated control.