In the current study, dark grayish-black sediments were collected from the east coast of Abu Dhabi-UAE (24°26′48.5″N; 54°26′40.6″E). Viviparous propagules of gray mangrove (A. marina) were obtained from either mother trees or freshly fallen ones collected from the same abovementioned location. Similar sized propagules were surface-sterilized using 70% ethyl alcohol (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) and 20% Clorox bleach. All surface-sterilized propagules were then washed ten times with 0.22-μm filter-sterilized (Millipore Corporation, Burlington, MA, USA) full-strength seawater (salinity of 40) and left to air-dry for 30 min.

In the current study, a commercial Acadian soluble SWE powder (Acadian Seaplants Limited, Dartmouth, NS, Canada) was used as a biostimulant. This organic seaweed concentrate is derived from the kelp, Ascophyllum nodosum. Physiochemical properties, macronutrients, and micronutrients of SWE can be found in Table S1 . Mangrove sediment was amended with 400 ml of SWE pot⁻¹ according to the manufacturer’s recommended rate (0.3 g SWE L⁻¹ water) every 14 days for the entire period of the in vivo experiments.

Mangrove propagules (Section 2.1) sown in plastic pots (23 cm diameter × 17 cm depth) containing sediment from the area described above were watered daily with full-strength seawater under greenhouse conditions (temperature of 25°C ± 2°C; relative humidity of 60% ± 5%; average daily photosynthetic photon flux density of 700 ± 150 μmol m⁻² s⁻¹). Eight pots (each containing two propagules) were prepared.

After 5 weeks, 16 seedlings from eight pots were collected and transferred to the laboratory. Roots were cut and washed, and fresh weight (FW) was recorded. Roots were soaked in sterile phosphate-buffered saline (PBS) solution (pH 7.0) for 10 min (Rennie et al., 1982) and surface-disinfested. For the surface sterilization, roots were first exposed to propylene oxide (Sigma-Aldrich) vapor for 25 min (Sardi et al., 1992). Roots were soaked in 70% ethyl alcohol for 4 min and 1.05% NaOCl for 4 min, followed by rinsing ten times in PBS.

Sterility checks were carried out for each sample before root maceration (Hallmann et al., 1997). The slurry was filtered through a sterile cotton cloth, and the filtrate was serially diluted (10⁻², 10⁻³, and 10⁻⁴). Aliquots (200 μl) were spread on plates containing inorganic salt starch agar (ISSA; Küster, 1959). For each root sample dilution, three replicated plates were dried for 15 min followed by 7-day incubation at 28°C ± 2°C in the dark (El-Tarabily et al., 2019). Population density (PD; log₁₀ colony-forming units (cfu) g root FW⁻¹) of endophytic actinobacteria was calculated (Hallmann et al., 1997). Colonies of streptomycete actinobacteria (SA) and non-streptomycete actinobacteria (NSA) were purified and identified on oatmeal agar plates supplemented with 0.1% yeast extract (OMYEA; ISP medium 3; Shirling and Gottlieb, 1966). Hyphae/spores of actinobacterial isolates were stored in 20% glycerol at −70°C (Wellington and Williams, 1977).Culture characteristics such as the color of aerial and substrate mycelia, and the production of diffusible pigments in addition to the presence or absence of aerial mycelia, the distribution of spores on both aerial and substrate mycelia, the formation of sporangia, and the stability/fragmentation of substrate mycelia were used to differentiate between SA and NSA (Cross, 1989). Filter-sterilized full-strength seawater was used in the preparation of all microbiological media in the current study.

All endophytic isolates were streaked on an ISSA medium supplemented with 80 g L⁻¹ (8%) of NaCl (Williams et al., 1972). Plates were incubated at 28°C in the dark for 7 days. Strong growth and heavy sporulation of actinobacterial isolates indicated high salt tolerance.

The high salt-tolerant SA and NSA were preliminarily tested for the production of indole-3-acetic acid (IAA). Briefly, 2 ml of each isolate (10⁸ cfu ml⁻¹) was incubated on a 250-rpm orbital shaker incubator in flasks containing 50 ml of inorganic salt starch broth (ISSB; Küster, 1959) supplied with 5 ml of 5% l-tryptophan (Sigma-Aldrich) for 7 at 28°C in the dark (Khalid et al., 2004; El-Tarabily et al., 2019). Suspensions were centrifuged at 12,000 × g, and 4 ml of the Salkowski reagent was added to the collected supernatants (Gordon and Weber, 1951). IAA-equivalents (µg ml⁻¹) were quantitatively determined using a spectrophotometer (UV-2101/3101 PC; Shimadzu Corporation, Analytical Instruments Division, Kyoto, Japan) at 530 nm.

All promising IAA-producing isolates were further grown for 10 days at 28°C in the dark in glucose peptone broth (di Menna, 1957) supplemented with 5 ml of 5% l-tryptophan to detect IAA and indole-3-pyruvic acid (IPYA) and on Strzelczyk and Pokojska-Burdziej (1984) medium to detect gibberellic acid (GA₃) and cytokinins (CKs) including isopentenyl adenine (iPa), isopentenyl adenoside (iPA), and zeatin (Z), using reverse-phase high-performance liquid chromatography (HPLC; SpectraLab Scientific Inc., Ontario, Canada). Auxins, GAs, and CKs were separated by using two isocratic solvent systems, according to Tien et al. (1979). Waters Associates HPLC with a differential ultraviolet detector was used to analyze the chromatograms, which were created by injecting 10 µl of the methanol dissolved extract onto a 10-m reverse-phase column (Waters Associates Bondapak C18, 4 mm × 30 cm) (El-Tarabily et al., 2020). In order to calculate the concentrations of PGRs in the unknown sample, their respective peak areas were compared with those obtained with authentic samples (Sigma-Aldrich) of a known concentration.

For ACCD assay, all isolates were plated on Dworkin and Foster’s (DF) salts minimal agar medium (Dworkin and Foster, 1958) supplemented with either 0.3033 g l⁻¹ of ACC (DF-ACC; Sigma-Aldrich) or 2 g l⁻¹ of ammonium sulfate (DF-(NH₄)₂SO₄; control) for 7 days at 28°C in the dark (El-Tarabily et al., 2019). Growth/sporulation on DF-ACC plates indicated that the isolate could produce ACCD. We also quantified the enzymatic activity of ACCD by measuring the amount of α-keto-butyrate (Honma and Shimomura, 1978). Protein concentrations were determined according to Bradford (1976).

To test the putrescine (Put) production, plates containing Moeller’s decarboxylase agar medium (MDAM) were supplemented with 2 g l⁻¹ of l-arginine-monohydrochloride (Sigma-Aldrich) and phenol red (Sigma-Aldrich) (Arena and Manca de Nadra, 2001). MDAM plates were incubated for 2 days in the dark at 28°C (El-Tarabily et al., 2020). The dark red halo surrounding the colonies indicated a Put-producing isolate. We also quantitatively test these Put-producing actinobacterial isolates for their production of the polyamines Put, spermidine (Spd), and spermine (Spm) using reverse-phase HPLC (Marino et al., 2000). Positive Put producers were placed in Moeller’s decarboxylase broth medium (MDBM) amended with 2 g l⁻¹ of l-arginine-monohydrochloride (Arena and Manca de Nadra, 2001). With the use of a 254-nm UV detector (Smith and Davies, 1985), an aliquot of 10 µl of the sample was injected onto a Bondapak C18 column (4 mm × 30 cm) in a liquid chromatograph (Waters Associates) as described by Marino et al. (2000).

For siderophore production, plates of chrome azurol S (CAS) agar (Schwyn and Neilands, 1987) were inoculated with isolates and incubated for 3 days at 28°C in dark. Actinobacterial isolates that were considered siderophore producers developed a yellow-orange halo zone around the colony.

Phosphate-solubilizing actinobacteria (PSA) were assayed using Pikovskaya (PVK) agar medium (Pikovskaya, 1948) supplemented with rock phosphate (Tianjin Crown Champion International Co. Limited, Tianjin, China) and amended with bromophenol blue. The appearance of a clear zone underneath the colony indicated a PSA isolate. The same isolates were also grown in 20 ml of sterilized National Botanical Research Institute Phosphate (NBRIP; Nautiyal, 1997) broth at 28°C for 2 days on a shaker at 150 rpm. For each isolate, aliquots (1 ml) were transferred to a flask containing 250 ml of NBRIP medium and incubated with continuous shaking at 28°C. Sterilized un-inoculated medium served as a control. After 3 days, a 10-ml sample of each culture or control was centrifuged at 12,000 × g for 15 min. The supernatant was used to determine the drop in pH and the amount of P released into the medium. The pH was recorded using a pH meter, whereas P availability was determined using the phospho-molybdate blue color method (Murphy and Riley, 1962).

Acetylene-reduction assays (Dye, 1962) and Nessler’s reagent (Holguin et al., 1992) were used to measure nitrogenase activity and NH₃ production, respectively. In all in vitro assays, eight independent replicates were used for each strain.

Of all PGP traits tested in vitro, the strongest PGPA isolates (#11, #12, and #20) were selected for the preliminary growth promotion experiment under gnotobiotic conditions. In a greenhouse, surface-sterilized mangrove propagules were sown in plastic pots containing sediment for 10 days and watered daily with full-strength seawater.

In order to prepare the inoculum for all the gnotobiotic experiments, 4 ml of aliquots of 20% glycerol suspension of the selected endophytes were individually inoculated into 250 ml of ISSB and shaken at 250 rpm on an orbital shaker incubator for 5 days. Cells were centrifuged at 12,000 × g for 15 min at 20°C, and the pellet was suspended in 10 ml of PBS and re-centrifuged (El-Tarabily et al., 2019). For each suspension, 0.1 ml of each of the dilutions of 10⁻³, 10⁻⁴, 10⁻⁵, and 10⁻⁶ was made in PBS and spread on ISSA. After 5 days of incubation, a final concentration of ~10⁸ cfu ml⁻¹ of each isolate was used as an inoculum.

Selected endophytic actinobacterial isolates were introduced inside the young seedlings using the pruned-root dip method (Musson et al., 1995). Briefly, root tips (3 mm) from germinated propagules were trimmed, and young seedlings were placed in sterile plastic cups containing the inoculum suspension (10⁸ cfu ml⁻¹) of each isolate for 3 h at 25°C. Seedlings of mangroves with or without the actinobacterial inoculum were aseptically planted into glass tubes (300 × 35 mm) filled with sediment and moistened with seawater. The control treatment was represented by seedlings that were placed in autoclaved ISSB.

All seedlings maintained in a growth chamber (day-time cycle, 16/8-h light/night; temperature, 25°C/20°C; fluorescent light, 180–200 µmol m⁻² s⁻¹) were irrigated daily with full-strength sterilized seawater. After 6 weeks of transplantation, plants were harvested, washed, and separated into roots and shoots. Dry weight (DW; g) and length (cm) of the shoot and root tissues were measured. Each treatment representing one seedling was independently replicated eight times.

Rifampicin resistant mutants of the promising PGPA (#11, #12, and #20) and non-PGPA positive control (#7) isolates were selected on ISSA medium supplemented with rifampicin (100 µg ml⁻¹; Sigma-Aldrich) and compared to the corresponding wild-type strains (Misaghi and Donndelinger, 1990). Features, such as morphology growth and PGP, of these mutants were found to be similar to the parental strains.

The pruned-root dip method was used to inoculate 10-day-old seedlings of mangroves with the endophyte inoculum (Section 2.4) in order to evaluate the colonization of internal root and stem tissues by isolates. Free-draining pots (36 cm in diameter) were filled with 14 kg of sediments (Section 2.1) and watered daily with full-strength seawater to container capacity in the greenhouse (Section 2.2). Roots and stems were sampled, washed, and surface-sterilized (Section 2.1) every 3 weeks (for 15 weeks) after planting. Samples were homogenized to determine the PD of isolates on ISSA amended with rifampicin. Each replicate represents a single pot containing one seedling, and each treatment was replicated eight times.

For light microscopy (LM) and transmission electron microscopy (TEM), specimens (6-week-old mangrove seedlings inoculated with the selected PGPA isolate) were fixed in freshly prepared Karnovsky’s fixative (2% paraformaldehyde (Sigma-Aldrich) + 2.5% glutaraldehyde (Sigma-Aldrich) in 0.17 M of Sorensen’s phosphate buffer, pH 7.2) for 24 h at 4°C. After being washed three times with the buffer, tissues were post-fixed with 1% aqueous osmium tetroxide for 2 h at 25°C, dehydrated with ascending grades of ethanol (30%–100%), and dipped into the propylene oxide (Sigma-Aldrich). Finally, samples were infiltrated, embedded in epoxy resin (Epon 812, Agar Scientific, UK), and polymerized at 60°C in an embedding oven for 24 h (Millonig, 1976). Blocks were trimmed into semi-thin sections (1.5 µm) and ultra-thin sections (95 nm) with Leica EM7 ultra microtome (Vienna, Austria). Slides of selected heat-dried, semi-thin sections stained with 1% toluidine blue and 1% borax (Sigma-Aldrich) were examined using Olympus BH-2 (Olympus Optical Co. Ltd, Tokyo, Japan) LM equipped with a digital camera and software (Jenoptik ProgRes Camera, C12plus, Frankfurt, Germany). For the TEM study, ultra-thin sections were collected on 200 mesh copper grids, stained with 10% uranyl acetate and 3% lead citrate, and examined using Tecnai Spirit G2 Biotwin TEM (FEI Co., Eindhoven, Netherlands).

The most potent isolate was identified based on the 16S rRNA gene sequencing, performed by Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ), Germany. Primers targeting 16S rRNA gene: 900R (5′-CCGTCAATT CATTTGAGTTT-3′), 800F (5′-ATTAGATACCCTGGTAG-3′), and 357F (5′-TACGGGAGGCAGCAG-3′) were used (Rainey et al., 1996; Saeed et al., 2017; Kamil et al., 2018). All 16S rRNA gene sequences of 16 representatives from the genus Streptomyces were retrieved from the National Center for Biotechnology Information (NCBI) database (http://www.ncbi.nlm.nih.gov/nucleotide/). Multiple sequence alignment of 16S rRNA genes was carried out using CLUSTAL-X (Thompson et al., 1997) implemented in the Molecular Evolutionary Genetics Analysis 7.0 (MEGA7) software with default parameters (Kumar et al., 2016). A phylogenetic tree of 16S rRNA genes was reconstructed using the maximum likelihood (ML) method (Felsenstein, 1981) implemented in the MEGA 7.0 software. Bootstrap values were calculated with 1,000 resamples.

The spore chain morphology and spore surface were examined using Philips XL-30 scanning electron microscopy (SEM; FEI Co., Eindhoven, Netherlands) of 15-day-old cultures grown on ISP-3 (Kumar et al., 2011).

In free-draining pots filled with sediments (Section 2.5), seedlings were inoculated with the most promising endophytic isolate #12 using the pruned-root dip method (Section 2.4). A total of four treatments were carried out: 1) control (seedlings inoculated with autoclaved ISSB medium only; neither SWE nor isolate was applied), 2) St (seedlings inoculated with endophytic isolate #12), 3) SWE (seedlings supplemented with only SWE), and 4) St+SWE (seedlings inoculated with endophytic isolate #12 in a sediment supplemented with SWE). SWE was added as per the manufacturer’s recommended rate (Section 2.1). Seedlings were watered daily with full-strength seawater to container capacity (Section 2.5). In the greenhouse, a randomized complete block design (RCBD) was used, where each replicate (total of eight replicates) was determined by a single pot containing one seedling. Trials were independently repeated twice with similar results.

The DW and length of roots and shoots, number of branches, and leaf surface area (cm²) were recorded at the end of 9 months post-planting (mpp) of propagules. Chlorophyll (Chl) fluorescence measurements were performed at 645 (for Chl a) and 663 nm (for Chl b) (Holden, 1965). Carotenoid pigments were measured at 470 nm according to Davies (1965).

The PAs (Put, Spd, and Spm) extracted from tissues of apical root and shoot tissues (Flores and Galston, 1982) were quantitatively determined using benzoyl chloride (Sigma-Aldrich) and normal internal standard of PAs (Redmond and Tseng, 1979). Reverse-phase HPLC chromatograms were produced onto a 10-µm reverse-phase column (Marino et al., 2000; Section 2.3).

To quantitatively measure the endogenous PGRs (auxins, CKs, GA₃, and abscisic acid [ABA]), crude extracts from mangrove root and shoot tissues (Shindy and Smith, 1975; Guinn et al., 1986; Machàckovà et al., 1993) were detected with the UV at 254 nm using reversed-phase HPLC (Waters Associates). This method provides quantification of phytohormones in a single run from 50 mg of fresh plant tissue. ACC content was assayed by the method of Lizada and Yang (1979). Derivatization of ACC was carried out by the addition of phenylisothiocyanate (Sigma-Aldrich), and the reverse-phase HPLC chromatograms were produced as described by Lanneluc-Sanson et al. (1986).

All measurements were taken on seedlings at the end of 9 mpp of propagules. Sixteen replicate samples from two independently repeated experiments were analyzed for all the tested parameters for each treatment.

We selected the tissues of the terminal part of the root and shoot systems at the end of 9 mpp of propagules, washed them in deionized water, and cut them into small pieces, which were then dried overnight at 70°C. Mineral nutrients were then analyzed in the roots and shoots, as follows: nitrogen was measured using a LECO FP-428 nitrogen analyzer by combusting finely ground plant material at 950°C in oxygen. As the sample passed through a thermal conductivity cell, the amount of N released from it was measured (Sweeney and Rexroad, 1987). Plant samples were digested in a 9:1 solution of nitric acid and perchloric acid to measure phosphorus (P), K, sulfur (S), magnesium (Mg), calcium (Ca), sodium (Na), Fe, manganese (Mn), copper (Cu), and zinc (Zn) (McQuaker et al., 1979). For the measurement of boron (B), plant material was dry-ashed and extracted with dilute acid, and B was colorimetrically measured with azomethine H (Gaines and Mitchell, 1979). Analyses were performed for all the nutrients in 16 replicates from two independently repeated experiments.

In free-draining pots filled with sediments (Section 2.5), seedlings were inoculated with the promising endophytic isolate #12 using the pruned-root dip method (Section 2.4). A total of four treatments were carried out as described above (Section 2.7). For each treatment, eight separate pots each containing one seedling were arranged in an RCBD. The container open-field nursery experiments were carried out on the Arabian Gulf coast of Abu Dhabi in the same location described in Section 2.1, and the nursery trials were independently repeated twice.

Control and inoculated seedlings were maintained under natural conditions between February and October (relative humidity range = 22%–43%; daytime length range = 11.0–13.5 h day; average temperature = 35.0°C ± 9°C day/23.0°C ± 8.0°C night; average precipitation = 6.7 mm). Irrigation was carried out naturally from the sea during the high and low tide times of the day. The DW and length of roots and shoots, number of branches, and leaf surface area (cm²) were recorded from 16 samples from two independently repeated experiments at the end of 12 mpp of propagules.

To study the effect of SWE on the growth of PGPA, including Streptomyces tubercidicus UAE1, a new SWE agar (SWEA) medium was developed. This medium was prepared by dissolving 40 ml of SWEA in 1 l of filter-sterilized full-strength seawater (pH 7.5). All endophytic isolates were streaked on SWEA, and the plates were incubated at 28°C in the dark for 7 days. Strong growth and sporulation of actinobacterial isolates on SWEA indicated the ability of SWE to support the growth of the endophytic isolates.

In order to compare the effect of SWE individually or in combination with S. tubercidicus UAE1 on the mangrove sediment ecosystem, total microbial activity was assessed at the end of the greenhouse and open-field nursey experiments. The microbial activity was measured using the fluorescein diacetate (FDA) hydrolysis technique (Schnurer and Rosswall, 1982). Briefly, 5 g of each sediment was added to 20 ml of 60 mM potassium phosphate buffer (8.7 g of K₂HPO₄ and 1.3 g of KH₂PO₄ in 1 l of distilled water, pH 7.6) in 250-ml flasks. The FDA was dissolved in acetone and stored as a stock solution (2 mg ml⁻¹) at −20°C. The reaction was started by adding 0.2 ml of FDA (400 µg) from the stock solution to a buffer–sediment mix. The reaction flasks were shaken (250 rpm) at 25°C for 20 min on an orbital shaker incubator. The reaction was then stopped by adding 20 ml of acetone to all samples. Sediment residues were centrifuged at 12,000 × g for 15 min and filtered using Whatman filter paper (Whatman, Maidstone, UK). The filtrate was collected in a test tube, covered with parafilm, and placed into an ice bath to reduce the volatilization of the acetone. The concentration of fluorescein was determined by reading the optical density at 490 nm using a spectrophotometer (Shimadzu Corporation). For each treatment, the background absorbance was corrected with the blank sample run under identical conditions but without the addition of FDA. Standard curves were prepared according to Chen et al. (1988). The results were converted to μg hydrolyzed FDA g dry sediment⁻¹. All analyses were collected from sediments at the end of 9 and 12 mpp of propagules of the greenhouse and open-field nursery experiments, respectively. Sixteen replicate samples from two independently repeated experiments were analyzed for each treatment.

In all experiments, RCBD was performed. Gnotobiotic greenhouse and nursery experiments were repeated with similar results. All data were combined and analyzed using the ANOVA procedure of SAS Software version 9 (SAS Institute Inc., NC, USA). Mean values of treatments were compared using Fisher’s protected least significant difference (LSD) test at p = 0.05 levels. The PD of the selected PGPA isolates was transformed into log₁₀ cfu g root or stem FW⁻¹.

The purified endophytic actinobacterial isolates were in vitro assessed for their salinity tolerance and different PGP traits. From the 26 endophytic actinobacteria, 20 SA and 6 NSA were isolated from the surface-sterilized root samples, and seven (6 SA and 1 NSA) isolates (#3, #11, #12, #17, #20, #23, and #25) were able to perform different PGP activities with an ability also to tolerate 8% NaCl ( Table 1 ; Figure S1 ). Regardless of their PGP traits, the rest of the isolates were not included in further experiments because they did not grow or sporulate on ISSA containing 8% NaCl.

We first tested the highly salt-tolerant isolates for the production of PGRs, ACCD, siderophores, nitrogenase, and NH₃ and for their ability to solubilize insoluble P. The identified isolates synthesized significantly (p < 0.05) different levels of auxins (IAA and IPYA), GA₃, and CKs (iPa, iPA, and Z) ( Table 1 ). Three isolates (#11, #12, and #20) produced higher amounts of auxins ( Table 1 ; Figure S1 ), whereas five (#11, #12, #20, #23, and #25) produced higher levels of GA₃ than any other tested isolates ( Table 1 ). Although isolate #3 produced the lowest levels of auxins and GA₃, isolate #17 did not produce GA₃ in the culture extracts. By comparing the in vitro production of CKs among the tested actinobacteria, isolate #12 produced the highest amounts of the three types ( Table 1 ). Thus, other isolates synthesized comparable amounts of iPa (#20) and iPA (#17) to isolate #12. Except for #3, #23, and #25, all other isolates produced significantly (p < 0.05) high levels of Z ( Table 1 ).

The production of enzymes ACCD and nitrogenase known for their PGP activities (El-Tarabily et al., 2019) by the actinobacterial isolates were also assessed in vitro. Although isolates #11, #12, #17, #20, and #25 produced the highest level of ACDD ( Table 1 ; Figure S1 ), none of these isolates produced nitrogenase ( Table 1 ).

We also qualitatively and quantitatively determined the production of PAs on culture extracts of the seven isolates. The amounts of Put, Spd, and Spm synthesized by these isolates significantly (p < 0.05) varied ( Table 1 ). By using the HPLC, isolates #11, #12, #20, #23, and #25 produced significantly (p < 0.05) higher levels of Put than the other tested isolates ( Table 1 ; Figure S1 ). Isolates #11, #12, and #20 synthesized the highest amounts of Spd and Spm ( Table 1 ). When we checked the production of siderophores and NH₃ in vitro, #17 and #25 were the only actinobacterial isolates that showed these PGP traits. The remaining produced either one or none of the two traits ( Table 1 ).

Six isolates (#3, #11, #12, #17, #20, and #23) were able to solubilize P ( Table 1 ; Figure S1 ). Since we tend to find the isolate(s) showing most, if not all, PGP activities, with the highest values in all the PGRs and ACCD tested, only isolates #11, #12, and #20 were further included for their endophytic existence and abundance in planta.

The strongest actinobacterial isolates showing multiple in vitro PGP traits were selected to determine the growth promotion of mangroves under gnotobiotic conditions. First, the chemical characteristics of the sediment collected earlier (Section 2.1) were analyzed and recorded as follows: pH (8.36), electrical conductivity (5.81 dS m⁻¹), organic carbon (6.24%), total P (85 and 8.83 mg kg⁻¹ sediment), available P (8.83 mg kg⁻¹), N (4 and 6.4 mg kg⁻¹ as nitrate and ammonium, respectively), bicarbonate extractable potassium (K; 241 mg kg⁻¹), amorphous iron (Fe) oxides (331 mg kg⁻¹), and sulfate (414 mg kg⁻¹).

In the greenhouse, endophytic isolates #11, #12, and #20 showed variable effects on mangrove growth. The three isolates significantly (p < 0.05) increased lengths and DWs of root and shoot as compared to non-inoculated (negative control) or inoculated with isolate #7 (positive control) treatments ( Table S2 ). The measured lengths of roots and stems and DWs of mangrove seedlings treated with the endophytic isolate #12 were significantly (p < 0.05) higher than those non-treated control or treated with isolates #11 or #20. Although plants inoculated with isolate #11 showed significantly (p < 0.05) increased lengths and DWs of shoots, we did notice the same effects on roots, compared to those inoculated with isolate #20.

To determine the longevity of the promising endophytic isolates in planta, we compared their PD in the internal tissues of mangrove roots and stems on a triweekly basis (up to 15 weeks). In our greenhouse experiments, isolates #11, #12, and #20 recovered in all tissues albeit at the sampling time. This suggests an endophytic nature of these PGPA isolates without causing harm to the host plant, i.e., mangrove. The total population of isolates, except for #11, increased significantly (p < 0.05) from the beginning until the end of the experiment in both tissues ( Figure 1 ). Following the significant (p < 0.05) increase in the first 6 weeks of colonization, there was a drop in PD of isolate #11 inside root and stem tissues of mangroves by the end of the experiment. We also noticed that the mean of the total population of isolate #20 in mangrove stems in Weeks 6–15 was insignificant (p > 0.05; Figure 1 ). Together, our data suggested that isolate #11 did not sufficiently recover from the tissues of mangrove seedlings after Week 6 and that isolate #20 inefficiently colonized stem tissues. Thus, the two isolates were excluded from further experiments. Our results imply, however, that #12 can be a potential PGPA isolate to be considered as a potential endophytic PGPA isolate. This was evident from the PD of this isolate that strikingly (p < 0.05) increased for the period of colonization in root and stem tissues throughout the greenhouse trials ( Figure 1 ). Compared to isolate #12, we noticed similar increase patterns in the PD of non-PGPA isolate #7 (control). Overall, the increased growth and high PD consistently found in seedlings by isolate #12 under controlled gnotobiotic and greenhouse conditions, respectively, made this isolate ideal for a large-scale experiment, i.e., a non-controlled open-field nursery.

To confirm the endophytic lifestyle of isolate #12, we used microscopic analyses to visually investigate the presence of spores and/or mycelia within the internal tissues of mangrove roots and stems at 6 weeks post-inoculation (wpi). With the use of LM, the spores of actinobacterial isolate #12 colonized the root tissues intercellularly within the parenchyma cells of the cortex and the xylem ( Figure 2A ). In addition, mycelial growth carrying the spiral spore chains of isolate #12 was detected within the cortical cells of roots ( Figure 2B ). Interestingly, the germinated spores and the formed germ tubes indicated the ability of isolate #12 to colonize the intracellular spaces of root cells ( Figure S2 ). Similarly, the endophytic isolate could successfully reside within the stem cells ( Figure 2C ). This was confirmed by the growth and survival of isolate #12 as part of its life cycle in vascular structures with its host plant ( Figure 2D ). The obtained results suggested that isolate #12 can translocate between the root and shoot systems through conductive tissues of the xylem.

With the use of TEM, the presence of the spore chains and substrate mycelia of isolate #12 was determined within the cortex of roots ( Figure 3A ) and stems ( Figure 3B ). We also figured out individual spores colonizing the cortical cells ( Figure 3C ) and attaching the cell membrane of host plant roots ( Figure 3C ) and stems ( Figure 3D ). This indicates a beneficial association between isolate #12 and mangrove seedlings. It is worth mentioning that we did not observe plant cell defects in any of the imaged samples in these microscopic studies. Overall, our data suggested that isolate #12 is an actinobacterial endophyte that inhabits the living root and stem tissues of mangroves.

Genomic DNA was extracted from the promising endophytic actinobacterial isolate, and the 16S rRNA gene was amplified (1,523 bp), sequenced, and deposited in GenBank under accession number MT883495. Next, the amplified fragment was used to perform a comparative sequence analysis with sequences available in GenBank. Our results identified actinobacterial species belonging to the genus Streptomyces. The identity between the 16S rRNA gene sequences obtained from isolate #12 and those available in GenBank ranged from 97.9% to 100.0%, of which Streptomyces nigrescens NBRC12894, Streptomyces libani NBRC13452, and S. tubercidicus NBRC13090 showed the highest similarities ( Figure 4A ). To distinguish it from other Streptomyces species, we described the pure cultures of this particular strain and the morphological characteristics of its spore chains. On ISP medium 3, the actinobacterial isolate developed light brownish gray mass color and yellowish-brown substrate mycelium with the production of yellow pigment on the reverse side of cultures ( Figure 4B ). The spore chains belonged to section Spirales, consisting of 3–10 mature, smooth-surfaced spores per chain ( Figure 4C ). Together, the molecular phylogeny, culture characteristics, and morphology of spores classified isolate #12 as S. tubercidicus Nakamura, 1961 Strain UAE1.

To determine their effect on growth promotion, mangrove seedlings were grown in sediment supplemented with SWE and/or inoculated with S. tubercidicus UAE1 (St) under greenhouse conditions. After SWE application or St inoculation, mangroves grew and developed into healthy plants ( Figure 5A ). The DW ( Figure 5B ) and length ( Figure 5C ) of roots and shoots apparently increased with either SWE or St compared to non-treated plants at the end of 9 mpp of the inoculated mangrove propagules, thus significantly (p < 0.05) varied between the two treatments. Seedlings of mangroves treated with the combination of SWE+St, however, showed the greatest DW and length of root and shoot systems. In line with that, the number of branches significantly (p < 0.05) increased by 40.6% and 55.3% in St-inoculated plants and SWE-treated, respectively ( Figure 5D ). We also noticed that the total leaf area was larger in plants treated with SWE by 29.0% or St by 18.3% than in non-inoculated plants ( Figure 5E ). The number of branches and total leaf area per plant were recorded at the highest values with SWE+St by 64.8% and 38.0%, respectively.

In planta photosynthetic pigments largely determine the photosynthetic capacity and hence plant growth (Li et al., 2018). Our results showed that contents of Chl a and Chl b pigments were significantly (p < 0.05) higher in SWE-treated plants than in non-inoculated or St-inoculated plants at 9 mpp in the inoculated mangrove propagules ( Figure 5F ). Among all the treatments, the combined treatment of SWE and St had the highest chlorophyll contents. Although there was no significant difference in the amounts of carotenoids in plants treated with SWE, St, or SWE+St, we observed that any of these treatments significantly (p < 0.05) stimulated the production of this particular photosynthetic pigment as compared to the control. Our data suggested that the growth of mangroves can be enhanced as a result of the photosynthesis and biomass production, more in plants treated with the combined treatment than the individual treatments of SWE or St.

In the greenhouse, we also studied the effect of SWE and/or St application on the endogenous contents of PGRs (phytohormones and ACCD) and mineral nutrients in roots and shoots at 9 mpp in the inoculated propagules associated with the growth promotion of mangroves. In general, all PGR levels investigated in this study were higher in shoot than in root tissues. The concentration of auxins (IAA and IPYA) in the tissues of plants grown in sediment supplemented with SWE, inoculated with St or both, were significantly (p < 0.05) different from the control treatment and each other ( Figure 6A ; Figure S3 ). Seedlings grown in sediment supplemented with SWE or inoculated with St were characterized by about 28.3%–30.6% or 16.7%–17.3% higher IAA concentration in root and shoot, respectively, than those in control treatments. IAA levels increased by 42% in root and 33% in shoot tissues in plants supplied with the combined two treatments. Compared to control, there were greater contents of IPYA in roots (22.3%–41.4%) and shoots (18.0%-28.0%) in seedlings supplied with any of the three treatments, of which SWE+St was the highest ( Figure 6A ; Figure S3 ). Similarly, the concentration of three types of CKs (iPA, iPa, and Z) varied significantly (p < 0.05) among all treatments in root tissues ( Figure S3 ). Although iPA and iPa contents increased significantly (p < 0.05) by SWE, St, and SWE+St treatments, we did not find significant (p > 0.05) differences in Z concentrations between control and St in shoots ( Figure 6B ). Treatments of SWE or SWE+St on mangrove seedlings increased Z in shoot tissues to similar levels, which were significantly (p < 0.05) higher than those with or without St ( Figure 6B ). There was no significant (p > 0.05) difference in the endogenous levels of ABA in mangrove root and shoot tissues in the four treatments ( Figure S4 ). Our data suggested that the growth promotion of mangroves can be increased by the application of SWE or inoculation of St by increasing the endogenous levels of auxins and CKs in tissues; thus, these can be greatly enhanced by SWE+St treatment.

In addition, the contents of GA₃ increased and ACC decreased significantly (p < 0.05) by St whether it was individually inoculated or combined with SWE in the examined tissues of mangroves ( Figures 6C, D ; S3 ). In general, there was a drop in ACC contents of 40%–50% in root and shoot tissues upon applying St. This suggests that the ACCD secreted by the local strain of S. tubercidicus UAE1 efficiently relieves plants from stress.

Under controlled conditions, there was a significant (p < 0.05) difference in the endogenous levels of Put, Spd, and Spm in tissues of all treatments ( Figures 6E ; S3 ). When St was inoculated or SWE was applied in pots containing seedlings of mangroves, roots had significantly (p < 0.05) higher levels of Put (34.2% or 53.2%, respectively), Spd (23.4% or 44.4%, respectively), and Spm (22% or 37.3%, respectively) than those grown without any biostimulant/bioinoculant. The treatment of SWE+St, however, increased Put by 62.7%, Spd by 61.1%, and Spm by 47.0% in the same tissue. Likewise, the pattern of increase in the three PAs was clearly demonstrated in shoot tissues as follows (from the highest to the lowest): SWE+St > SWE > St > control ( Figures 6E ; S3 ).

In addition, we examined if other growth-promoting substances (e.g., macro- and micronutrients) of St and SWE have a role in plant growth and development. In general, we found significant (p < 0.05) differences in the tissue contents of all examined macro- and micronutrients in plants treated with SWE ( Table 2 ). We also noticed that there could be an additive effect of SWE when provided simultaneously with St. For instance, SWE or St significantly (p < 0.05) increased N in both tested tissues as compared to control plants ( Table 2 ). Thus, these three nutrients increased to the highest levels when SWE and St were applied together compared to their corresponding individual treatments. Notably, there was a significant (p < 0.05) increase in the measured available P in sediments and contents of P and K in roots and shoots in plants treated with SWE and St+SWE compared to those non-treated (control) or inoculated with St only ( Table 2 ). This suggests that SWE, but not St, can positively regulate nutrient (e.g., P) availability in sediments, thus increasing P and other nutrients in mangrove tissues for improved growth characteristics.

In addition, there were significant (p < 0.05) differences in the concentrations of S, Mg, Ca, Na, B, Fe, Mn, Cu, and Zn in the shoots and roots of SWE- or St+SWE-treated plants as compared to those non-inoculated or inoculated with St ( Table 2 ), although all the measured macro- and micro-elements were insignificantly (p > 0.05) different between control and St treatments. This suggests that St does not have a direct effect on the uptake of these mineral nutrients. In general, St+SWE-treated plants showed a comparable effect in the concentrations of the abovementioned elements in tissues compared to those SWE treated. Our data indicated that the biostimulant SWE and the bioinoculant S. tubercidicus UAE1 can compensate for the lack/deficiency of mineral nutrients in sediments and regulate endogenous PGRs in planta, thus enhancing the growth of mangroves and improving the efficiency of photosynthesis.

The DW and length of roots and shoots were clearly increased with either St or SWE compared to non-treated plants at 12 mpp of the inoculated mangrove propagules under non-controlled open-field nursery conditions; thus, these parameters significantly (p < 0.05) varied between the two treatments ( Table 3 ). Mangrove plants treated with St+SWE demonstrated, however, the greatest effect on their DW and length. Along with that, the number of branches significantly (p < 0.05) increased by 1.7 and 2.2 times in St-inoculated and SWE-treated plants, respectively ( Table 3 ). In addition, the total leaf area was larger in plants treated with St by 21.6% or SWE by 38.1% times than in non-inoculated plants. The number of branches and total leaf area per plant were recorded as the highest with St+SWE, reaching threefold and 63.1%, respectively.

Endophytic PGPA isolates grew and sporulated on SWEA. This suggests that SWE can serve as a nutrient/food base for the growth of PGPA isolates without showing any adverse effect on the growth and multiplication of isolates. Microbial activity in sediment amended with SWE was found to be significantly (p < 0.05) higher than in the non-amended control sediment in both the greenhouse and open-field nursery experiments ( Table S3 ). An increase of 66.4% after 9 mpp in the greenhouse and 75.3% at 12 mpp was observed in the open-field nursery of mangrove seedlings supplemented with SWE compared to the control sediment. Thus, there was no significant (p > 0.05) difference between SWE treatments, with or without inoculation with St. When comparing St-inoculated and non-inoculated (control) sediments under controlled (i.e., greenhouse) and non-controlled (i.e., open-field nursery) conditions, no significant (p > 0.05) difference was detected in the microbial activities ( Table S3 ). This suggests that SWE can increase the number of soil microbiota, in addition to its growth benefits when applied to marine plants, such as mangroves.