In all experiments, wild-type seeds of S. bigelovii Torr. (Scrops, Ninove, Belgium) were surface sterilized by 70% ethyl alcohol (EtOH; Sigma-Aldrich Chemie GmbH, Germany) and 1.05% Clorox (20% household bleach). Two drops of Tween 20 (Sigma-Aldrich) was used in all surface-sterilization steps. Surface-sterilized seeds were then washed with 0.22 μm membrane filter-sterilized (Millipore Corporation, MA, United States) full strength seawater (salinity of 4.0%), and were air-dried before use.

Pale grayish-yellow sandy soil was collected from Al Rams coast, Ras Al Khaimah, UAE (25° 52′ 44″ N, 56° 1′ 25″ E), where S. bigelovii is naturally growing. Soil was passed through a 1 cm mesh sieve for all greenhouse experiments. The following soil chemical characteristics were detected: electrical conductivity 9.31 dSm^–1; pH 8.18 (in 0.01 M CaCl₂), organic C 1.15%, and the nutrients (mg kg soil^–1) [N (5.3 as NH₄⁺; 2.9 as NO₃^–), oxalate extractable amorphous Fe (382), sulfate (311), bicarbonate extractable K (265), total P and available P (44, 8.3, respectively)].

Rhizosphere soils of six young S. bigelovii seedlings were collected from the above-mentioned study site. Excavated soil (maximum depth of 15 cm) from around the root balls was placed into sterile plastic bags and brought to the laboratory. Samples were passed through 5 mm sieve of which each sample consisted of rhizosphere soils from two seedlings, randomly pooled as one replicate. The three collected rhizosphere soil samples were air-dried at 25°C for 4 days to reduce the viable vegetative bacterial cells (Williams et al., 1972).

The aerobic culturable actinobacterial populations of the freshly sampled rhizosphere soils were determined using the soil dilution plate count method. Three replicates of 10 g rhizosphere soil samples were dispensed into 100 mL of membrane filter-sterilized full strength seawater, and the soil suspension was placed in an ultra-sonic cleaner at a frequency of 55,000 cycles s^–1 for 20 s. The soil suspension was then shaken at 250 rpm at 28°C for 30 min on a rotary shaker, Model G76 (New Brunswick Scientific, NJ, United States).

Dilutions of 10^–2–10^–5 were made in membrane filter-sterilized full strength seawater and 0.2 mL aliquots were spread onto inorganic salt starch agar (ISSA) (Küster, 1959) made with membrane filter-sterilized full strength seawater in sterile petri dishes. Cooled (45°C) sterile ISSA medium was amended with 50 μg mL^–1 cycloheximide (Sigma-Aldrich) and 50 μg mL^–1 nystatin (Sigma-Aldrich). Five plates per dilution were used and air-dried in a laminar flow.

Plates were incubated for 7 days at 28°C in dark and colonies were counted (log₁₀ colony forming units (cfu) g^–1 dry soil). Unless otherwise indicated, actinobacteria were purified and maintained on ISP3 medium (oatmeal agar) plates made with membrane filter-sterilized full strength sweater and supplied with 0.1% yeast extract (OMYEA) (Shirling and Gottlieb, 1966). Streptomycete actinobacteria (SA) and non-streptomycete actinobacteria (NSA) were tentatively identified to the genus level according to Cross (1989). Colonies of SA and NSA were identified based on morphological characteristics, ability to form sporangia, distribution of aerial and/or substrate mycelia, presence or absence of aerial mycelia, the stability and fragmentation of substrate mycelia (Cross, 1989). Isolates were stored as mycelia and spores in 20% glycerol at −20°C (Wellington and Williams, 1977).

Tolerance to salt stress was evaluated by growing the selected isolates on ISSA medium made with deionized water and supplemented with the following NaCl concentrations: 0, 10, 20, 40, 60, and 80 g L^–1 medium. The isolates were streaked in triplicates on the plates and incubated for 7 days at 28°C in dark (Williams et al., 1972). Actinobacterial growth and heavy sporulation on ISSA supplemented with 80 g NaCl L¹ (8%) indicated the efficiency of the selected isolates to tolerate high NaCl concentration and to be true halotolerant isolates. Only these isolates were chosen for subsequent experiments.

The ability of the obtained halotolerant actinobacterial isolates to colonize S. bigelovii was carried in vitro using the primary qualitative root colonization plate assay (Kortemaa et al., 1994) to determine if S. bigelovii root exudates can act as the only carbon source to support the growth of each isolate. Surface-sterilized S. bigelovii seeds were pre-germinated on membrane filter-sterilized full strength seawater agar plates for 2 days before being transferred onto new plates (one seed/plate). Eight replicates were inoculated with each isolate grown on OMYEA (one isolate/plate). Seeded plates without treatment were used as controls. Actinobacterial suspension (approximately 10⁸ cfu mL^–1) was prepared for each isolate from OMYEA plates made with membrane filter-sterilized full strength seawater. An inoculum (0.2 mL) was placed 1–2 mm alongside the emerging radicle (Kortemaa et al., 1994). The seawater agar plates were vertically incubated at 28°C in dark and the root colonization by the isolates was detected after 8 days. Colonization was calculated as a percentage of the total root length (Kortemaa et al., 1994).

According to the indicator root colonization plate assay, roots were initially fixed with Karowskys fixative (2.5% glutaraldehyde and 2% para-formaldehyde) for 4 h at 25°C. Samples were treated with 0.2 M phosphate buffer, pH 7.2 at 4°C for 2 h and post-fixed in 1% osmium tetroxide for 1–2 h. After washing with deionized water, samples were dehydrated with a series of EtOH dilutions ranging from 30 to 100% at 4°C. Once in 100% EtOH, roots were dried in a critical-point dryer (Polaron CPD Bell Brook Business Park, Bolton Close England) and mounted on carbon-tabbed aluminum (Al) stubs. Stubs were fixed in the Polaron sputter coater vacuum chamber and sputtered with gold Au/Pd target for 5 min at 20 mA. Phillips XL-30 SEM (Eindhoven, The Netherlands) was used to examine the colonized root samples.

All microbiological media in the experiments described in sections Production of IAA, PA, and ACCD by Actinobacteria and Other PGP Activities of the Three Promising Isolates were prepared using membrane filter-sterilized full strength seawater. The selected root colonizing isolates were in vitro tested for their capability to produce IAA, PA, and ACCD. To detect IAA, Erlenmeyer flasks (100 mL) containing 50 mL of sterile inorganic salt starch broth (ISSB) (Küster, 1959) were supplied with 5 mL of 5% filter-sterilized L-tryptophan (Khalid et al., 2004). The flasks were inoculated with 2 mL of each isolate prepared from a 5 days old shaken ISSB culture of about 1 × 10⁸ cfu mL^–1, covered with Al-foil and incubated on a rotary shaker (Model G76) at 250 rpm at 28°C in dark for 7 days. Non-inoculated flasks were used as a control. After incubation, suspensions from each flask were centrifuged at 12,000 × g for 30 min. The supernatant was filtered through Millipore membranes and collected in sterile tubes. Six milliliters of culture supernatants were pipetted into test tubes and 4 mL of Salkowski reagent (2 mL of 0.5 M FeCl₃ + 98 mL 35% HClO₄) were applied (Gordon and Weber, 1951). Tubes were left for 30 min to develop red color; and the intensity of the color was determined by optical density at 530 nm using a scanning spectrophotometer (UV-2101/3101 PC; Shimadzu Corporation, Analytical Instruments Division, Kyoto, Japan). A standard curve was established according to the color of the standard solutions of IAA, and auxin compounds were expressed as IAA-equivalents (Gordon and Weber, 1951). Eight replicate samples were analyzed.

The obtained actinobacterial isolates were tested for the production of putrescine (Put) as a polyamine using Moeller’s decarboxylase agar medium (MDAM) amended with 2 g L^–1 L-arginine-monohydrochloride (Sigma-Aldrich) and 0.02 g L^–1 phenol red (Sigma-Aldrich) (Arena and Manca de Nadra, 2001). Isolates growing on OMYEA were streaked in triplicate on MDAM plates with or without Arginine (control). Plates were incubated at 28°C for 3 days in dark. Production of Put by the decarboxylating isolates was determined by the existence of a dark red halo beneath and around the colonies.

All isolates showing large red halo color on MDAM were further tested for production of Put, spermidine (Spd), and spermine (Spm) in Moeller’s decarboxylase broth medium (MDBM) amended with 2 g L^–1 of L-arginine-monohydrochloride by reverse-phase high-performance liquid chromatography (HPLC; SpectraLab Scientific Inc., ON, Canada). Erlenmeyer flasks containing 50 mL of MDBM were inoculated with 3 mL of 20% glycerol suspension of each isolate (10⁸ cfu mL^–1) and incubated on a rotary shaker at 250 rpm at 28°C. After incubation for 10 days in dark, the suspension from each flask was centrifuged at 12,000 × g for 30 min. The membrane filter-sterilized cell-free culture broth was collected in sterilized McCartney tubes.

The derivatization of the filter-sterilized cell-free filtrates and the internal polyamine standards were carried out as follows: 0.5 mL of filtrate was added to 0.5 mL of saturated solution of sodium bicarbonate (Sigma-Aldrich) and 2 mL of dansyl chloride (Sigma-Aldrich) solution (10 mg mL^–1 in acetone). The reaction vessel was incubated at 60°C for 20 min and 50 μL of 28% ammonium hydroxide were added. After 30 min, the reaction mixture was filtered again through 0.22 μm Millipore membranes and used for the HPLC (Marino et al., 2000). The HPLC chromatograms (eight replicates of each isolate) were produced by injecting 10 μL aliquot of the sample onto a 10 μm μBondapak C₁₈ column (4 mm × 30 cm) in liquid chromatograph (Waters Associates) equipped with a 254 nm UV detector (Smith and Davies, 1985).

The selected isolates were also screened for the production of ACCD. Five-day-old isolates grown on rich OMYEA were streaked in triplicates on N-free Dworkin and Foster’s salts minimal agar medium (DF) plates (Dworkin and Foster, 1958) amended with either 3 mM ACC (Sigma-Aldrich) or 2 g (NH₄)₂SO₄ (control). The heat-labile ACC was filter-sterilized and the filtrate was added to the salt medium. Plates were incubated at 28°C in dark for 7 days. Isolates growing on DF + ACC plates indicated the efficiency of the isolate to utilize ACC and produce ACCD. To determine the activity of ACCD, isolates were grown in ISSB at 28°C for 5 days in dark. Cells were centrifuged, washed with 0.1 M Tris-HCl (pH 8.5) and inoculated onto DF + ACC broth on a rotary shaker at 250 rpm for another 5 days. Cells were resuspended in 0.1 M Tris-HCl and ruptured by freezing/thawing cycles thrice (Shah et al., 1998). The lysate was centrifuged at 80,000 × g for 1 h. The enzymatic activity of ACCD was assayed by monitoring the amount of α-ketobutyrate liberated (Honma and Shimomura, 1978). Protein concentrations were determined as described (Bradford, 1976). Eight replicate samples were analyzed.

To investigate whether particular members of actinobacteria producing high levels of IAA, PA, or ACCD are avid colonizers of the rhizosphere (rhizosphere-competent), the non-sterilized soil tube assay was used (Ahmad and Baker, 1987) using rifampicin resistant mutants (Misaghi and Donndelinger, 1990).

The surface-sterilized S. bigelovii seeds were coated with each isolate as described below. Non-treated seeds served as controls. In each tube containing non-sterilized field soil (collected from the same site described above), one seed was planted 5 mm deep. Tubes were vertically placed into boxes containing the same soil and watered with full strength seawater to container capacity. Boxes were covered and held upright with no more added seawater to the tubes after sowing (Ahmad and Baker, 1987). Boxes were kept in a greenhouse (photosynthetic photon flux density of 700 μmol m^–2 s^–1) at 25 ± 2°C and relative humidity (RH) of 60 ± 5%.

After 4 weeks, S. bigelovii were harvested from the soil tube assay. The length of the harvested roots was measured, and only the top 12 cm of roots were retained, and were aseptically cut into 2 cm segments (Ahmad and Baker, 1987). Loose particles of the rhizosphere soil on root segments were removed and air-dried. Rhizosphere soil particles were numbered based on the root segments from which they were recovered (Ahmad and Baker, 1987). Roots and soil particles were separately studied to determine the root-colonization frequencies (R%) and population densities (PD; log₁₀ cfu g^–1 dry soil) of each isolate in the rhizosphere soil, respectively (Ahmad and Baker, 1987) on ISSA plates made with membrane filter-sterilized full strength sweater and supplemented with 100 μg mL^–1 rifampicin (Sigma-Aldrich). To be considered as colonized-rhizosphere or -root segment, colonies of the isolated strains were obtained on ISSA plates amended with rifampicin either in the corresponding rhizosphere soil sample, on the root segment or both (Ahmad and Baker, 1987) after 7 days of incubation at 28°C in dark. Percentage of root segments colonized was calculated as the proportion of the total number of segments colonized to the total segments sampled at each distance. Experiments were repeated twice with eight replicates in each; and only the strongest three rhizosphere-competent isolates producing either IAA, PA, or ACCD were chosen for the greenhouse experiments.

Amplification of 16S rRNA gene using polymerase chain reaction (PCR) of the three isolates (#7, #1, and #21) and their sequencing were carried out by Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ, Braunschweig, Germany). According to Rainey et al. (1996), the PCR primers used were: 900R (5′-CCGTCAATTCATTTGAGTTT-3′); 357F (5′-TACGGGAGGCAGCAG-3′) and 800F (5′-ATTAGATACCCTGGTAG-3′). Nucleotide sequences were submitted to GenBank and assigned accession numbers MN795132, MN795133, and MN795131 for isolates #7, #11, and #21, respectively.

Pairwise sequence similarity using 16S rRNA gene sequence was determined. For the phylogenetic analyses, reference strains were selected from the top hits of the determination using GenBank BLAST¹. An amplification of 1,516, 1,519, and 1,519 bp from isolates #7, #11, and #21 was obtained and aligned with closely-related strain sequences of Streptomyces using CLUSTAL-X (Thompson et al., 1997) in Molecular Evolutionary Genetics Analysis 7.0 (MEGA7) software (Kumar et al., 2016). Phylogenetic trees were constructed using the maximum likelihood (ML) method of which bootstrap values were calculated based on 1000 resamplings. The morphology of the spore chains and surface was carried out for the three isolates using Phillips XL-30 SEM.

To detect auxins as IAA and IPYA (indole-3-pyruvic acid), the three isolates (#7, #11, and #21) were cultivated in 50 mL ISSB supplemented with 5 mL of 5% L-tryptophan (Sigma-Aldrich). Gibberellic acid (GA₃) and cytokinins [isopentenyl adenine (iPa), isopentenyl adenoside (iPA) and zeatin (Z)], were detected on Strzelczyk and Pokojska-Burdziej (1984) medium. After 10 days, the extraction of these PGRs from the concentrated filter-sterilized cell-free broth and the HPLC parameters used to determine the concentrations of IAA, IPYA, GA₃, iPa, iPA, and Z were carried out as described by Tien et al. (1979).

The production of siderophores was determined using chrome azurol S agar plates (Schwyn and Neilands, 1987). Yellow-orange halo zone developed around the colony was considered positive for siderophores production. The N-fixing activities of isolates were examined by the acetylene reduction assay (Holguin et al., 1992) using Varian 6000 gas chromatogram (Varian Instrument Group, United States). Production of NH₃ was determined using Nessler’s reagent (Dye, 1962). Solubilization of insoluble rock phosphate (Tianjin Crown Champion International Co., Ltd., Tianjin, China) was estimated using Pikovskaya’s agar medium (Pikovskaya, 1948) with bromophenol blue in which tricalcium phosphate was replaced with rock phosphate. Production of clear zone was an indicator of P-solubilization. All rock phosphate-solubilizing actinobacterial isolates showing large zones of clearing on Pikovskaya’s agar were further tested quantitatively for their abilities to solubilize rock phosphate in modified National Botanical Research Institute’s phosphate broth (Nautiyal, 1997) containing 5 g L^–1 rock phosphate. Water-soluble P was analyzed by the colorimetric procedure of Murphy and Riley (1962) using molybdophosphoric acid blue complex. The drop in pH and the released soluble P amount were taken as an indicator of the efficiency of isolates. In all the above-mentioned tests, eight independent replicates were used for each isolate.

The spot-inoculated plate technique described by de Boer et al. (1999) was used to determine the isolate’s sensitivity to the diffusible metabolites of the other isolates. Suspensions of actinobacterial spores were cultured on OMYEA in 10 mM MgSO₄. For each isolate, approximately 10⁸ cfu mL^–1 was used as an inoculum. Actinobacterial isolates were spot-inoculated on OMYEA by pipetting 2 droplets of 5 μL of the actinobacterial inoculum; and plates were incubated at 28°C for 5 days in dark. Later on, a suspension of the target isolate was atomized over the surface of the spot-inoculated plates. After 4 days at 28°C in dark, zones of growth inhibition of the target isolate around the spot-inoculated isolates were evaluated, if any.

To determine if the selected isolates do not inhibit each other by their volatile compounds, the method described by Payne et al. (2000) was used. Briefly, fresh OMYEA plates were inoculated by evenly spreading 0.2 mL of 10⁸ cfu mL^–1 of the isolate onto the surface of agar. Plates were incubated in dark at 28°C. After 5 days of incubation, another fresh plate of OMYEA was simultaneously inoculated with an actively growing culture of the tested isolate. The lids were removed and the plates containing the tested isolate were inverted over the previously grown isolate; and the two plate bases were taped together with Parafilm (American National Can TM, Greenwich, CT, United States). Control plates were prepared the same way except that non-inoculated plate was used instead of a plate containing the target isolate. After another 5 days in incubator, the growth of the target isolate was compared to that of the control. In these two tests, four independent replicates were used for each isolate.

For the preparation of activated carbon as seed adsorbents for the actinobacterial isolates to be used in greenhouse experiments, local date seeds from the fruits of date palm (Phoenix dactylifera) were dried and ground using an electric agitated mortar (JK-G-250B2, Shanghi Jingke Scientific Instrument) (Mathew et al., 2018). Physical activation of date seeds was performed in a tube furnace (GSL-1500X; MTI Corporation, VA, United States) with carbonization followed by activation. A carbonization step was started by allowing N₂ gas to pass through the furnace for 10 min when the temperature was gradually increased under a constant flow of N₂ at a rate of 5°C min^–1 up to 600°C and maintained at this temperature for 4 h (Sekirifa et al., 2013). The carbonaceous material was activated at 900°C in the same furnace under the flow of CO₂ gas. The activated carbon was then sieved using a testing mesh (size of 200–300 μm). To prepare the activated carbon nanomaterials, activated carbon was wet-ground in a grinder (Retsch RM 100, Germany) and dried in a freeze dryer (Telstar, Spain) at −55°C and 0.02 mbar for 6 h (Mathew et al., 2018). The material was then filtered using a 0.45 μm polytetrafluoroethylene filter (Thomas Scientific, NJ, United States) prior to use as a seed adsorbent for actinobacterial isolates.

Inocula of each isolate were applied as seed coating and in soil for the greenhouse experiment described below as recommended by El-Tarabily (2008).

For the seed coating, isolates grown in ISSB made with membrane filter-sterilized full strength sweater were shaken at 250 rpm at 28°C for 5 days in dark. Cells and/or spores were collected and centrifuged for 20 min at 5,000 × g. Pellets were washed three times before being suspended in sterile 0.03 M MgSO₄ (buffer). Instantly, surface-sterilized S. bigelovii seeds were immersed in beakers containing either each actinobacterial suspension (approximately 10⁸ cfu mL^–1) in 0.03 M MgSO₄ (buffer) or buffer alone (no actinobacterial suspension; control) for 4 h at 28°C. Activated carbon from date seeds (5 g L^–1) was added as an adsorbent for the treated and the control seeds. The rest of the suspension was drained and seeds were air-dried.

For each isolate, inocula were prepared for the soil application by adding 300 g of oat bran as a food source in Erlenmeyer flasks (500 mL) autoclaved for 20 min at 121°C on three successive days. The oat bran was then aseptically inoculated with spore suspensions (25 mL) of each isolate in 20% glycerol (10⁸ cfu mL^–1) and incubated in dark at 28°C for 3 weeks. Non-colonized oat bran, which had been autoclaved twice, served as a control. Colonized oat bran with each isolate (0.05 g colonized oat bran inoculum g^–1 air-dried non-sterile soil) were mixed thoroughly using a cement mixer in all experiments described below.

S. bigelovii growth was in vivo tested to determine the effect of the three isolates #7, #11, and #21 either individually or combined three together under greenhouse conditions. Briefly, free draining plastic pots (23 cm in diameter) were filled with 7 kg soil sampled from the area (described above) and mixed with 0.05 g of colonized oat bran inoculum g^–1 field soil. Surface-sterilized healthy S. bigelovii seeds were coated with each isolate and with the activated carbon from date pits (described above) or with activated carbon only (control). In each pot, eight seeds were sown in soil at 5 mm deep, and seedlings were thinned to four/pot when emergence was complete. A total of five treatments were applied; where each treatment was replicated eight times with four seedlings/replicate as follows: Non-inoculated control (C); individually inoculated with either isolate #21 (Sc), isolate #7 (St) or isolate #11 (Sr), producing auxins, PA or ACCD, respectively; and collectively inoculated with all three isolates (Sc/St/Sr).

Pots were placed in the greenhouse (photosynthetic photon flux density of 700 μmol m^–2 s^–1) at 25 ± 2°C and RH of 60 ± 5%) in a randomized complete block design (RCBD), and were daily watered to container capacity with full strength seawater. Dry weight and length of shoots and roots were recorded at 12 weeks post-sowing (wps) the seeds, and seed dry weight was recorded 20 wps.

The levels of photosynthetic pigments, including chlorophyll (chl a and chl b) and carotenoids, were determined in the succulent stems (Holden, 1965; Davies, 1965). Tissues of the terminal part of the root and shoot systems were used for the analysis of endogenous auxins, PA, and ACC. The free PA (Put, Spd, and Spm) were extracted according to Flores and Galston (1982). Benzoylation using benzoyl chloride (Sigma-Aldrich) and the HPLC for the plant extracts and the internal standard of PA were carried out as described by Redmond and Tseng (1979). The extraction of endogenous IAA, IPYA was carried out as previously described (Shindy and Smith, 1975; Guinn et al., 1986). The HPLC parameters were applied as previously described (Tien et al., 1979). The extraction of endogenous ACC was carried out according to Lizada and Yang (1979). Derivatization of ACC was done by adding phenylisothiocyanate (Sigma-Aldrich) and the HPLC chromatograms were produced as described (Lanneluc-Sanson et al., 1986). Eight replicate samples were analyzed for the detection of photosynthetic pigments and endogenous auxins, PA, and ACC. All procedures were carried out on seedlings at 12 wps.

Treatments were arranged in a RCBD in all experiments. Data from the experiments were repeated with similar results, combined and analyzed. Data of PD were transformed into log₁₀ cfu g^–1 dry soil, and R% were arc-sine transformed before analysis of variance (ANOVA) was carried out. Data were subjected to ANOVA and means were compared using Fisher’s Protected LSD Test (P = 0.05).

For multidimensional analyses, a PCA was performed to determine the contribution of PGP to the experimental variance. Results were displayed in a biplot. The PCA was carried out using XLSTAT version 2019.3.2 (Addinsoft, United States). Pearson’s correlation was conducted between all pairs of studied variables within the same treatment of actinobacterial isolates and plant tissues. For all statistical analyses carried out in this study, SAS Software version 9 was used (SAS Institute Inc., NC, United States).

The density of actinobacteria populations in S. bigelovii rhizosphere was 5.36 ± SE 1.33 log₁₀ cfu g^–1 dry soil. Thirty-nine actinobacteria were isolated from the rhizosphere on ISSA plates (Supplementary Figure S1). According to their cultural and morphological characteristics, 27 (69.23%) and 12 (30.76%) were identified as SA and NSA isolates, respectively. All Streptomyces spp. and genera of the NSA (Actinomadura, Actinoplanes, Microbispora, Micromonospora, Rhodococcus, Nocardia, Nocardiopsis, and Streptosporangium spp.), were enumerated (Supplementary Figure S1) and purified on OMYEA plates.

Eventhough all 39 isolates grew well and sporulated heavily on ISSA medium supplied with up to 40 g L^–1 NaCl, only 22 (56.41%) showed increased tolerance to 80 g L^–1 of NaCl (Supplementary Figure S1). This indicates the efficiency of the selected isolates to tolerate high concentration of NaCl (8%) and were considered to be true halotolerant isolates. Of the 22 halotolerant isolates, 13 (59.1%) and (9) (40.9%) isolates belonged to SA and NSA, respectively.

In order to determine the ability of the halotolerant actinobacterial isolates to colonize S. bigelovii roots, an in vitro indicator root colonization plate assay was performed. Eight of halotolerant actinobacterial isolates (#1, #3, #10, #14, #23, #29, #33, and #37) failed to colonize the roots of S. bigelovii at 8 days after emergence. Therefore, they were excluded from the subsequent studies. The remaining 14 isolates (#2, #4, #6, #7, #9, #11, #15, #18, #21, #25, #27, #31, #32, and #35) showed different degrees of root colonization after 8 days of radicle emergence; 10 isolates (#2, #4, #6, #7, #11, #18, #21, #25, #27, and #32) revealed 100% of root colonization, whilst the rest (#9, #15, #31, and #35) colonized 50–75% of the roots. The latter four isolates were excluded in further studies. Only the 10 isolates showing whole root colonization (7 SA and 3 NSA) were selected.

The full colonization of the 10 isolates for S. bigelovii root with plate assay was confirmed by using SEM. For example, it was evident that the actinobacterial isolates #7, #11, and #21 that fully colonized the root system when evaluated by the in vitro root colonization plate assay. They also showed extensive mycelial growth and spore chains on the root surface of S. bigelovii with the use of SEM (Figure 1).

Using the colorimetric analysis, auxins were detected in the liquid cultures of isolates #4, #18, #21, #27, and #32 (Table 1). The efficiency of these isolates in their capability for IAA production varied greatly. The other five isolates did not produce IAA. The isolates that formed dark red color after the addition of Salkowski reagent were considered as IAA-producing isolates (Supplementary Figure S2).

We also screened the isolates for the ability to produce Put on MDAM plates amended with L-arginine-monohydrochloride. The relatively moderate to large dark red halo around and beneath the colonies in isolates #2, #4, #7, #25, and #32 indicated the production of Put (Table 1 and Supplementary Figure S2): but the absence of red halo was an indicator of inability of the remaining isolates to produce Put. From the HPLC analysis of the culture extracts of the 10 tested isolates grown on MDBM amended with L-arginine-monohydrochloride, the production of the PA (Put, Spd, and Spm) varied significantly (P < 0.05) (Table 1).

Five out of the 10 isolates grew and sporulated on DF + ACC agar (Supplementary Figure S2); the rest, however, grew only on DF control medium with (NH₄)₂SO₄ (Table 1). Quantitative estimation of ACCD activity revealed significant (P < 0.05) variations in the production of ACCD among the different isolates (Table 1). We noticed that isolates #2, #4, #6, #11, and #27 produced moderate to high levels of ACCD activity and were considered ACCD producers for further analyses.

The rhizosphere competence assays were carried out to evaluate the root colonization abilities under a naturally competitive greenhouse environment for the 10 isolates that initially colonized the entire root system in vitro. There was significant (P < 0.05) variation in the root colonizing abilities of the different isolates (Table 2). The root-colonization frequencies of actinobcaterial strains in the rhizosphere soil were highest in isolates #7, #11, #21, and #27 that successfully colonized both the root and the rhizosphere up to 12 cm of root length (Table 2). Although the 10 isolates colonized the rhizosphere soil at all soil depths, their abundancies were significantly (P < 0.05) greater at the top 8 cm in comparison to deeper roots (Table 2). All 10 isolates colonized the root system of S. bigelovii and the rhizosphere soil (up to 12 cm depth), but isolates #11, #7, #21, and #27 showed the highest colonized PD values.

The strongest three rhizosphere-competent isolates #21, #7, and #11 that also were good producers for auxins, PA and ACCD, respectively, were selected to study their effects on S. bigelovii growth under greenhouse conditions. Notably, isolates #2, #4, #6, #18, #25, and #32 produced high levels of auxins, PA and/or ACCD (Table 1). However, none of the isolates were selected for the greenhouse experiments because they showed low levels of rhizosphere competency (Table 2). Similarly, isolate #27 was eliminated due to its multiple modes of action albeit the excellent rhizosphere competency. In this study, the criteria used in choosing the isolates were (1) to solely possess a single mechanism of action such as producing high levels of auxins, PA or ACCD; and (2) to be able to show a strong rhizosphere competency as recommended by El-Tarabily et al. (2020).

The promising isolates that only produced either auxins, PA or ACCD isolates were further identified and characterized. Purified PCR products of the 16S rDNA gene was sequenced. The resulting sequence data of isolates #21 (Genbank accession number: MN795131), #7 (MN795132), and #11 (MN795133) were compared with representative 16S rRNA gene sequences of organisms belonging to the actinobacteria. Based on the evolutionary distance values, the phylogenetic tree was constructed according to the neighborhood joining method (Saitou and Nei, 1987). The 16S rRNA gene of isolates #21, #7, and #11 were compared with sequences in the GenBank, European Molecular Biology Laboratory (EMBL) or Ribosomal Database Project (RDP) database (Maidak et al., 1999), which showed that these actinobacterial candidates were all belonged to the genus Streptomyces spp. The complete 16S rRNA gene sequence of isolate #21 showed full 100.0% similarity with S. chartreusis (NR 041216), although the remaining isolates of Streptomyces spp. showed less than 99.3% similarities (Figure 2A). On ISP3 medium, pure cultures of isolate #21 produced powdery blue-green aerial mycelium with yellow substrate mycelial growth after 7 days of incubation (Figure 2B and Supplementary Table S1). According to our observations, the isolate showed branched substrate mycelia from which aerial hyphae developed in the form of open spirals (Figure 2C). The configuration of the spore chains of isolate #21 which belonged to spirales section of 10–25 mature, spiny spores per chain was detected (Figure 2C and Supplementary Table S1). Our data suggest that isolate #21 can be recognized as Streptomyces chartreusis Leach et al. (1953) strain UAE1.

The phylogenetic analysis of isolate #7 showed 99.6% similarity to S. tritolerans DAS 165 (NR 043745) and S. tendae ATCC (NR 025871) and NBRC 12822 (NR 112290); while the rest showed < 99.2% similarity with this particular strain (Figure 3A). This suggests that this isolate could be S. tritolerans, S. tendae or a strain of a new species within the genus Streptomyces. The pure cultures produced dark yellow substrate mycelia and gray aerial mycelia when isolate #7 was incubated for 7 days on ISP3 medium (Figure 3B). The isolate did not produce any diffusible pigments in ISP3 medium, and it grew on plates of 7% of NaCl or more (Supplementary Table S2); all of these distinctive characteristics are common features of S. tritolerans (Syed et al., 2007). Spores of the isolate were smooth, oval-rod shaped (Figure 3C); and its spore chains arranged in long straight to flexous (rectiflexibiles) chains of 10–15 spores (Figure 3C and Supplementary Table S2). Our data further support the identification of isolate #7 can be recognized as Streptomyces tritolerans Syed et al. (2007) strain RAK1.

For isolate #11, it was phylogenetically grouped in the same clade as five other species, namely S. geysiriensis NRRL B-12102 (NR 043818), S. vinaceusdrappus BRC 13099 (NR 112368), S. plicatus NBRC 13071 (NR 112357), S. enissocaesilis NRRL B-16365 (NR 115668), and S. rochei NRRL B-1559 (NR 116078) with 100% similarity (Figure 4A). To identify our isolate from the very closely related species, detailed morphological, cultural, biochemical and physiological characterization were carried out to distinguish the ACCD-producing isolate. Light grayish-yellow color of aerial mycelium and pale grayish-yellow color of substrate mycelium were produced by isolate #11 on ISP3 medium without production of any diffusible pigments (Figure 4B). Except of S. enissocaesilis, it was difficult to distinguish between our isolate and the other four species based on the aerial mass or substrate mycelium color on any of the ISP media (Supplementary Table S3). At high magnification of SEM, the strain clearly developed smooth, oval-rod shaped spores forming straight to flexuous long chains (rectiflexibiles) of 20–35 spores per chain (Figure 4C). The other Streptomyces strains, however, showed either retinaculum-apertum or spiral spore chains (Supplementary Table S3). The identification was further confirmed when other phenotypic, biochemical and physiological characteristics between isolate #21 and S. rochei NRRL B-1559 (NR 116078) (Kämpfer, 2012) were comparable (Supplementary Table S3); suggesting that isolate #11 can most probably be Streptomyces rochei Berger et al. (1953) strain RAMS1. Together, our data indicate that the auxins, PA, and ACCD-producing isolates were identified as S. chartreusis, S. tritolerans and S. rochei, respectively.

The production of the PGRs, ACCD and siderophores, and P-solubilization abilities varied among S. chartreusis, S. tritolerans, and S. rochei (Table 3). Only the auxin-producing isolate S. chartreusis produced high levels of IAA and IPYA (44.35 and 9.11 μg mL^–1, respectively), but did not produce Put, Spd, Spm, or ACCD. Conversely, the PA-producing isolate S. tritolerans (#7) only produced high levels of Put, Spd, Spm (Tables 1, 3), but did not produce detectable levels of IAA, IPYA or ACCD (Tables 1, 3). Unlike the other two strains, the ACCD-producing isolate S. rochei (#11) produced only ACCD without the production of detectable levels of IAA, IPYA, Put, Spd, or Spm (Tables 1, 3). We did not notice detectable levels of gibberellic acid in the culture extracts of any of the three isolates (Table 3). There was a variation in the production of cytokinins by the three isolates. Isolates #7 and #11 produced 1.12 and 1.32 μg mL^–1 of iPa, respectively, and isolates #21 and #11 produced 0.95 and 1.52 μg mL^–1 of iPA, respectively. The two isolates #21 and #7 produced 1.45 and 1.05 μg mL^–1 of Z, respectively (Table 3).

The three tested isolates neither fixed N nor produced ammonia. Only two of the isolates (#7 and #21) were able to solubilize P and only one isolate (#21) was able to produce siderophores (Table 3 and Supplementary Figure S2).

Interestingly, the growth of S. chartreusis, S. tritolerans, and S. rochei did not appear to be affected by the presence of each other; there was no evidence of inhibitory effects of the volatile metabolites on the isolates’ growth. In addition, the sprayed target isolates were not inhibited by the spot-inoculated isolate, indicating no observed inhibitory effect of the diffusible metabolites on the growth of the target isolates.

In order to determine the effect of the three actinobacterial isolates as producers for specific PGP, we treated S. bigelovii with each isolate individually and in combinations (i.e., as a synergistic consortium). In general, the application of S. chartreusis (Sc), S. tritolerans (St), and S. rochei (Sr) either singly or in combination (Sc/St/Sr) significantly (P < 0.05) promoted growth of S. bigelovii compared to the control treatment after 12 wps of sowing S. bigelovii seeds (Figure 5A). Compared to the non-inoculated control plants, Sc treatment increased shoot and root dry biomass by 32.3 and 42.3%, respectively (Figure 5B). When seawater-irrigated plants were treated with either St or Sr, there was a significant (P < 0.05) increase in the dry biomass of shoots (46.9% in St and 56.5% in Sr) and roots (62.4% in St and 71.9% in Sr) compared to the control treatment. Similar observations were found in shoot and root tissues when we measured their lengths. When we treated S. bigelovii seedlings with Sc, St or Sr, we found that the length of both shoots and roots increased significantly (P < 0.05) compared to control (Figure 5C). Thus, Sc/St/Sr revealed the greatest growth. This was confirmed by the significant increases in the dry weights of shoots and roots by 62.2 and 77.9%, respectively (Figure 5B), and the length of shoots and root by 56.0 and 56.8%, respectively (Figure 5C) compared to non-inoculated control plants.

Treatments of S. bigelovii with the three individual isolates (Sc, St, or Sr) or with the combined approach (Sc/St/Sr) had significantly (P < 0.05) increased levels of chl a, chl b and carotenoids as compared with control plants (Figure 5D). The combined effect of the three isolates (Sc/St/Sr) resulted in more production of photosynthetic pigments compared the application of each isolate individually. Similarly, S. bigelovii yielded more seeds of 46.1% in Sc, 60.0% in St and 69.1% in Sr than that in the control plants watered with seawater at the harvest time (Figure 5E). Such increase in seed yield reached to 79.7% in the consortium treatment of Sc/St/Sr when compared to the control plants. In individual treatments, growth enhancement reflected as a set of morpho-physiological parameters (dry biomass, photosynthetic pigments and seed yield), was most pronounced with the ACCD-producing S. rochei (Figure 5). This was followed by the PA-producing S. tritolerans and lastly by the auxin-producing isolate S. chartreusis compared to non-inoculated control plants. Interestingly, the consortium had the superiority in growth promotion effect in comparison with any other individual applications. This suggests that the consortium of actinobacterial isolates stimulated growth of S. bigelovii, most probably, due to the synergistic interactions between these isolates via multiple mechanisms.

To study possible mechanisms leading to the growth enhancement of S. bigelovii after individual or combined treatments of isolates, we analyzed the endogenous levels of particular PGRs in shoot and root tissues of S. bigelovii. The application of Sc or Sc/St/Sr significantly (P < 0.05) increased the levels of IAA and IPYA in the tissues of S. bigelovii compared to control plants (Figures 6A,B). Plants treated with either the St or the mixture of isolates significantly (P < 0.05) increased the levels of endogenous Put, Spd and Spm than any other treatment including the control in both roots and shoot tissues (Figures 6C–E). Similarly, S. bigelovii treated with Sr or Sc/St/Sr significantly (P < 0.05) reduced the levels of ACC in the plant tissues compared to the control treatment (Figure 6F).

Application with the PA-producing isolate S. tritolerans (St) or the ACCD-producing S. rochei (Sr) significantly (P < 0.05) increased levels of endogenous auxins of S. bigelovii compared to control (Figures 6A,B). Both S. tritolerans and S. rochei were, however, found to be non-auxin-producing isolates. This suggests that the production of PA and ACCD by these isolates may indirectly increase the auxins levels in planta.

In general, the overall endogenous levels of auxins and PA measured in both tissues of S. bigelovii were found to be high, but low levels of ACC in Sc/St/Sr treatment were detected compared with other treatments (Figure 6). This suggests that the synergistic consortium of isolates positively regulated the activities of the endogenous hormone auxin and the PA with a concomitant decline in the level of ACC in planta more than those in the individual treatments. This ultimately could lead to better growth and performance of S. bigelovii. To lesser extent, the effects were also observed by the individual application of Sc, St or Sr.

The PCA was performed to identify the principal components (PCs) to identify the most important morpho-physiological parameters that are responsible about the variation of both shoot and root growth and seed yield of S. bigelovii that best describe the growth promotion to the actinobcaterial isolates individually or in combination. The first two PCs accounted for 60.8 and 17% of the total variation in shoots (77.8%; Figure 7A) and 56 and 20.5% of the total variation in roots (76.5%; Figure 7B). Plants received the five different treatments were segregated into clear five clusters; plants that were not treated with any actinobacterial isolate (control) were on side, but plants inoculated with the combination of the three actinobcaterial isolates (Sc/St/Sr) on the other side that had higher agronomic growth parameters associated positively with all PGRs in plant tissues. Plants treated with the auxin-producing S. chartreusis are located next to control plants followed by plants treated with ACCD-producing S. rochei. The relatively slower growth and seed yield of Sc and Sr treated plants, as compared to those treated with St and Sc/St/Sr were not associated with insignificant productions of Put, Spd, Spm and ACC. In fact, ACC was negatively associated in roots of plants treated with Sc and both roots and shoots of plants treated with Sr. The Scree plots representing the eigenvalue of variance explained by each PC in shoot and root were also shown (Supplementary Figure S3).

The effects of auxins, PA and/or ACCD produced by individual or combined isolates were also evaluated on plant growth parameters in shoot and root tissues. On the biplots, Sc, St, and Sr were closely located to auxins, PA, and ACCD, respectively (Figure 7). The consortium of isolates, on the other hand, was closely related to all growth parameters and PGRs. In general, the accumulation of growth parameters was positively correlated upon individual or mixed inoculations of isolates, but was negatively correlated with the endogenous ACC content of shoot and root. This suggests that each strain of actinobacteria possessing different mechanisms can directly contribute to plant promotion; thus the application of the consortium showing rhizopshere-competent abilities maximizes the growth and yield of S. bigelovii.

We also studied the association across the morpho-physiological and biochemical characteristics of S. bigelovii and the PGP produced by each isolate. The correlation coefficients (r-values) analysis showed that the auxin-producing S. chartreusis had significant (P < 0.05) correlation with all agronomic growth parameters, IAA and IPYA; but insignificant (P > 0.05) with the endogenous PA and ACC in both plant organs (Supplementary Tables S4, S5). Thus, 45 and 28 insignificant (P > 0.05) correlations with Sc in shoot and root, respectively. There were 14 and 10 insignificant (P > 0.05) correlations in shoots and roots of S. bigelovii inoculated with the PA-producing S. tritolerans, respectively (Supplementary Tables S6, S7); and 24 and 23 insignificant correlations in the same tissues with the ACCD-producing S. rochei treatment (Supplementary Tables S8, S9). Neither Sc nor St significantly correlated with ACC. Sr negatively correlated with ACC in planta. It is noteworthy to mention that most endogenous PA in shoot and root tissues were insignificant (P > 0.05) in S. bigelovii tissues after Sr treatment. This was confirmed when the three isolates were combined to determine the correlation coefficients between plant growth attributes in shoots and roots of S. bigelovii and endogenous growth regulation compounds. The synergistic consortium significantly (P < 0.01) correlated with almost all agronomic growth parameters and PGRs in plant tissues; and thus the insignificance was not observed between growth regulation parameters in root (Supplementary Tables S10, S11). This indicates that the promoting effects of the ACCD, three PA followed by the auxins detected in S. rochei, S. tritolerans, and S. chartreusis, respectively, were effective in enhancing different levels of growth regulating compounds. Thus, the relative superiority in performance of the consortium can be attributed to the multiple modes of action that are known to enhance growth and relieve plants from environmental stresses.