Halotolerant biofilm-producing rhizobacteria mitigate seawater-induced salt stress and promote growth of tomato

Biofilm-producing rhizobacteria (BPR) enhance productivity and mitigate abiotic stresses in plants. This study showed that 21 out of 65 halotolerant rhizobacteria could build biofilms. The components of the biofilm matrices i.e., extracellular polymeric substances (EPS) are proteins, curli, nanocelloluse, nucleic acids, lipids, and peptidoglycans. Various functional groups including carbonyl, carboxyl, amino, hydroxyl, and phosphate were identified. Positions of these groups were shifted by application of 5% NaCl, suggesting Na+ biosorption. By sequencing, Glutamicibacter arilaitensis (ESK1, ESM4 and ESM7), G. nicotianae (ESK19, ESM8 and ESM16), Enterobacter ludwigii (ESK15, ESK17, ESM2 and ESM17), E. cloacae (ESM5 and ESM12), Exiguobacterium acetylicum (ESM24 and ESM25), Staphylococcus saprophyticus ESK6, Leclercia adecarboxylata ESK12, Pseudomonas poae ESK16, Bacillus subtilis ESM14, and P. putida ESM17 were identified. These rhizobacteria exhibited numerous plant growth-promoting (PGP) activities including producing IAA, ACC deaminase, and siderophores, and solubilizing phosphate. Under non-stress, bacterized plants increased biomass accumulation (8–23.2% roots and 23–49.4% shoots), while under seawater-induced salt stress only ESK12, ESM4, ESM12, and ESM14 enhanced biomass production (5.8–52.9% roots and 8.8–33.4% shoots). Bacterized plants induced antioxidant defense system (19.5–142% catalase and 12.3–24.2% DPPH radical scavenging activity), retained a greater relative water content (17–124%), showed lesser membrane injuries (19.9–26.5%), and a reduced Na+ (6–24% in roots) and increased K+/Na+ ratio (78.8 and 103% in roots by ESK12 and ESM24, respectively) than the non-bacterized plants in saline conditions. Thus, native halotolerant BPR can be utilized as ameliorators of salt stress.


Results
Screening of halotolerant biofilm-producing rhizobacteria. A total of 65 halotolerant rhizobacterial isolates were isolated from 13 rhizosphere samples (5 isolates sample −1 ). Biofilm formation by rhizobacteria is one of the important features for ameliorating salinity stress. Thus, all (65) these halotolerant isolates were evaluated for their capacities to form air-liquid (AL)-or solid-air-liquid (SAL) biofilms on the flat bottom glass test tube containing salt-optimized broth plus glycerol (SOBG) media supplemented with 5% NaCl. Among them, 20 isolates (30.7%) built the AL biofilms, while only one isolate (ESM14) formed the SAL biofilm after 72 h incubation in a static condition at 28 °C (Fig. 1A). The remaining isolates neither generated AL nor SAL biofilms even after 2 weeks of incubation (data not shown). Biomass biofilm weight in terms of wet basis was quantified (Fig. 1B). The wet biomass biofilm weight was significantly (P ≤ 0.001) varied in these bacterial isolates, and ranged from 3.65 to 86.2 mg mL −1 . The highest biomass biofilm (86.2 mg mL −1 ) was noted in ESM7 which was statistically similar to ESK17 (85.55 mg mL −1 ) and ESM24 (84.84 mg mL −1 ). Incredibly, the lowest biomass biofilm was obtained in ESM14 (3.65 mg mL −1 ) which was not significantly differed from ESM25 (3.75 mg mL −1 ). Like the production of biomass biofilm, the CFU (colony forming unit mL −1 ) also differed significantly (P ≤ 0.001) in the different biofilm matrices, varying from 4.2E+08 to 9.85E+11 (Fig. 1C). The highest CFU was recorded in ESM24 (9.85E+11) though not significantly different from ESM7 (9.6E+11). The second highest CFU was noted in ESM5 (8.5E+11) followed by ESK17 (4.9E+11) and ESM23 (4.7E+11). The lowest CFU (4.2E+08) was found in ESM14 and ESM25.
Identification of halotolerant biofilm-producing rhizobacteria. Because of their negative results in the hemolytic tests, 19 out of 21 biofilm-producing rhizobacteria were sequenced. The isolates ESK1, ESM4 and ESM7 were identified as Glutamicibacter arilaitensis; ESK19, ESM8 and ESM16 as G. nicotianae; ESK15, ESK17, ESM2 and ESM17 as Enterobacter ludwigii; ESM5 and ESM12 as E. cloacae; and ESM24 and ESM25 belonged to Exiguobacterium acetylicum. The isolates ESK6, ESK12, ESK16, ESM14, and ESM17 were recognized as Staphylococcus saprophyticus, Leclercia adecarboxylata, Pseudomonas poae, Bacillus subtilis and P. putida, respectively. The 16S rRNA gene sequence data were deposited in the NCBI (National Center for Biotechnology Information) GenBank, and the allotted accession numbers are listed in Table 1. A phylogenetic tree was also constructed using the 16S rRNA gene sequences as presented in Fig. 2.

Detection of biofilm matrix components and their functional groups. The biofilm matrix/EPS
components of the identified rhizobacteria were not characterized yet. SEM (scanning electron microscope) was www.nature.com/scientificreports/ used to characterize the morphology of the biofilm matrices of the selected bacterial isolates. Numerous ribbonlike structures, popularly known as nanocellulose fibers 53 were observed in the biofilm matrices of S. saprophyticus ESK6, L. adecarboxylata ESK12, G. arilaitensis ESM4, E. cloacae ESM5, P. putida ESM17 and E. acetylicum ESM24. These fibers were highly interconnected and compact. Representative images are illustrated in Fig. 3a-c. FTIR (Fourier transform infrared) spectroscopy was used to identify the biomolecules and chemical functional groups/ligands present in the biofilm matrices. All the examined biofilm matrices consisted of proteins, polysaccharides, nucleic acids, lipids, and peptidoglycans (Fig. 3d). The peaks of amide I (1600-1700 cm −1 ) and amide II (1500-1600 cm −1 ) were remarkably higher than the peaks of polysaccharides (900-1150 cm −1 ), nucleic acids (1220-1250 cm −1 ), and lipids (2800-2970 cm −1 ). Several functional groups were identified based on available literatures [36][37][38] . The broad band around 3400 cm −1 was related to the O-H stretching mode of the hydroxyl Table 1. Identification of halotolerant biofilm producing rhizobacteria isolated from coastal regions of Bangladesh.

Isolates
Geographical position Sequence length (bp) Maximum score Identity (%) www.nature.com/scientificreports/ group, while absorptions at 2800-2970 cm −1 were assigned to the asymmetrical C-H stretching vibration of the aliphatic CH 2 -group. The bands at 1520-1650 cm −1 were attributed to the stretching vibration of C=O, N-H and C-N groups in the presence of proteins, while the band around 1400 cm −1 was assigned to C=O symmetric stretching of COO − . The bands at 1220-1250 cm −1 were characteristics for P=O stretching of PO 2 − phosphodiesters. The peaks at 900-1150 cm −1 were correlated to the C-O or C-C stretching vibrations, P=O stretching, and C-O-H or C-O-C deformation vibrations.
Congo red-and Calcofluor (a cellulose-specific dye) binding assays were used to determine whether biofilm producing bacterial isolates produce curli and/or cellulose or not. In this study, all the tested bacterial isolates formed the typical RDAR (red, dry, and rough) phenotype on plates containing Congo red (Fig. 3e) associated with assembly of cellulose and curli 54 . When these isolates were speckled onto the Calcofluor agar plates, they displayed a weak to strong fluorescence (Fig. 3f), suggesting these isolates produced cellulose 55 . However, cellulose production might vary between these bacterial isolates.  www.nature.com/scientificreports/

Expression of PGP traits in vitro.
IAA is one of the most studied phytohormones secreted by numerous rhizobacteria that stimulate plant growth. IAA production in the presence of 0.2% L-tryptophan was quantified using both Salkowsky colorimetric-and HPLC (High-Performance Liquid Chromatography) methods (Table 2). Except for ESK1 and ESK16, all other cultures developed the pink color after adding the Salkowsky reagent, suggesting a positive auxin production (data not shown). By spectrophotometry, IAA was not detected from nether in ESK1 nor in ESM16. IAA production ranged from 1.56 to 117.47 µg mL −1 in the colorimetric method. In this case, the maximum IAA was produced by ESM19 (117.47 µg mL −1 ) which was statistically similar with ESM2. All the isolates were found to synthesize IAA when quantified using HPLC. A higher amount of IAA was detected in all the isolates by HPLC than by the colorimetric method. IAA production ranged from 19.60 to 384.00 µg mL −1 in the HPLC method. In this case, ESM19 synthesized the highest amount of IAA (384 µg mL −1 ) followed by ESM5 (364.01 µg mL −1 ) and ESM2 (337.19 µg mL −1 ). The representative chromatographs of known concentration of IAA, ESM4, ESM8 and ESM19 are shown in Fig. 4a-d.
In order to prove the isolates were diazotrophs, the nifH gene was amplified by polymerase chain reaction (PCR) using nifH-specific primers. Except for ESK17 and ESK19, all the isolates amplified 470 bp and 359 bp nifH fragments ( Supplementary Fig. S1). To confirm the PCR results, the isolates were grown on N-free media. ESK17 and ESK19 did not form any colonies on N-free solid LG agar plates, in contrast to the other isolates. Thus, the majority of the isolates are diazotrophs. P solubilizing rhizobacteria screened using tricalcium phosphate was found ineffective to solubilize P under field conditions 56 . Hence, P solubilization in terms of P solubilization index (PSI) was compared using both tricalcium phosphate (0.5%) and rock phosphate (0.8%) as a P source (Table 2). Except for ESK19, ESM8 and ESM25, all other isolates were able to solubilize P from both sources. However, ESM16 was incapable to solubilize P from tricalcium phosphate but can solubilize P from rock phosphate. The PSI value varied remarkably (P ≤ 0.001) in the different isolates, and fluctuating between 2.07-7.25 and 1. 84-7.20 in tricalcium phosphate, and rock phosphate, respectively. Five isolates (ESK15, ESK17, ESM12, ESM19 and ESM24) were capable to solubilize K from potassium aluminum silicate, and only ESM24 solubilized Zn from Zn 3 (PO 4 ) 2 but not from ZnO or ZnCO 3 (Table 2).
Abiotic stress tolerance and antagonistic activities. The results of the abiotic stress tolerance analysis are illustrated in Supplementary Table S1. All the isolates grew at 10% NaCl except ESK17, whereas 14 out of the 19 isolates propagated at 15% NaCl on Luria Bertani (LB) agar plates. All the isolates also grew at pH 9 and pH 10. Except for ESK6, ESK12, ESM2, and ESM7, all other isolates grew at the acidic condition of pH 4. Testing of heat tolerance, except for ESK17, all other isolates grew at 42 °C. Interestingly, 14 out of the 19 isolates were able to even grow at 50 °C.
Biofilm-producing PGPR reduce seawater-induced biochemical changes in tomato plants. Electrolyte leakage and relative water content. In order to detect the effects of biofilm-producing PGPR isolates to protect cells from oxidative damages, electrolyte leakage (EL) and relative water content (RWC) were measured. After 8 days of seawater-induced salt stress, all the bacterized plants exhibited lesser EL than the non-bacterized plants (Fig. 6a) was not effective in reducing EL. Taken together, the association of PGPR with roots is an effective mechanism to protect from membrane leakage under normal and stressed conditions.
With an increased EL, the amount of RWC will be decreased. In this study, the leaf RWC was higher in all the PGPR-inoculated plants under seawater-induced salt stress conditions from 17 to 24% compared with the non-inoculated control plants (Fig. 6b). However, the RWC of the plants was not significantly affected by the inoculation of the PGPR isolates under non-stress conditions (Fig. 6b).  www.nature.com/scientificreports/ www.nature.com/scientificreports/ Accumulation of hydrogen peroxide and malondialdehyde. As an oxidative stress indicator, the accumulation of hydrogen peroxide (H 2 O 2 ) and malondialdehyde (MDA) in the non-stress and seawater-induced salt stress plants was measured. The non-inoculated stressed plants accumulated 12 ROS-oxidized membrane lipids generate MDA as a product of lipid peroxidation. Consistent with the reduction of H 2 O 2 , bacterized tomato plants also showed a reduced MDA accumulation under seawater-induced salt stress conditions (Fig. 6d). Although most of the PGPR isolates reduced MDA levels, ESM12 and ESM24 inoculation resulted in a more efficient reduction of MDA accumulation than in non-inoculated control plants. Bacterized non-stressed plants also exhibited reduced MDA levels. These suggest that biofilm-producing PGPR effectively reduces stress indicators even under normal growth conditions, though the indicators are increased in the seawater-induced salt stress conditions.
The protein content is also an indicator of the stress status of the plants. To prove the efficacy of biofilmproducing PGPR to reduce the seawater-induced salt stress, the total protein content was measured (Fig. 6e). The amount of total protein was significantly reduced in seawater-induced salt stress plants. All the bacterized plants reduced the degradation of total protein content except ESM12. ESM17, ESK12, and ESM4 decreased the protein degradation by 42, 30 and 26%, respectively, while ESK17, ESM14, and ESM24 reduced the total protein though insignificantly. Under non-stress conditions, bacterized plants also prevented the protein degradation, but it was not pronounced in the non-inoculated plants.
Antioxidants and radical-scavenging capacity of PGPR. Seawater-induced salt stress plants accumulated higher amounts of H 2 O 2 and MDA ( Fig. 6c,d). Thus, it is expected that free radical scavenging-and catalase (CAT) activity will also be increased in bacterized plants to scavenge H 2 O 2 and MDA for preventing oxidative damage. Under seawater-induced salt stress, CAT activity was increased twofold in plants inoculated with ESK17, ESM12, ESM17, and ESM24 as compared to non-inoculated plants, and 1.5-fold with ESK12 and ESM4 (Fig. 6f). Under non-stress, bacterized plants also increased the CAT activity.
Accumulation of anthocyanins in different parts of the plants is common under various adverse conditions. The antioxidant properties of the anthocyanins protect plants from various biotic and abiotic stresses including salt stress in plants 58 . In this study, non-inoculated seawater-induced salt stress plants accumulated the 13.53 µg anthocyanin g −1 FW. Except for ESM17, all the bacterized plants reduced (twofold) the anthocyanin contents (Fig. 6h). Anthocyanin accumulation was also visible in the leaves of the non-inoculated seawater-induced salt stress plants (Fig. 6i). Thus, biofilm-producing PGPR isolates are effective in attenuating the seawater-induced salt stress in tomatoes due to accumulation of anthocyanins.
Chlorophyll synthesis and biomass accumulation by application of PGPR. Inoculation of halotolerant biofilmproducing rhizobacteria decreased the accumulation of H 2 O 2 and MDA through an increased enzymatic activity under seawater-induced salt stress conditions in tomatoes ( Fig. 6f-h). Hence, it is anticipated that the inoculation of different biofilm-producing rhizobacterial isolates will also increase chlorophyll (Chl) content and biomass accumulation in tomatoes. Non-inoculated tomato plants under seawater-induced salt stress conditions had 1.01 mg g −1 FW Chl a. ESK12, ESM12, ESM17, and ESM24 but not ESK17, ESM4, and ESM14 increased the production of Chl a (ca. 12%) compared to the non-inoculated control plants (Fig. 7a). Similarly, all the examined isolates also boosted the synthesis of Chl b (Fig. 7b). However, only ESK12-and ESM14-inoculated plants significantly increased (32% and 28%, respectively) Chl b synthesis and to a lower extent also for ESK17 and ESM4 with 13% and 22%, respectively.
In general, higher chlorophyll contents positively affect biomass accumulation in plants. In this study, ESK12and ESM12-inoculated plants significantly increased the biomass accumulation in both roots and shoots under seawater-induced salt stress conditions compared to the non-inoculated control plant by about 30% (Fig. 7c) and 50% (Fig. 7d), respectively. ESM14-and ESM4-inoculated plants also showed increased biomass accumulation in both roots and shoots under seawater-induced salt stress conditions. On the other hand, ESK17-and ESM24inoculated plants were only slightly increased in shoot biomass at the early growth stage of tomatoes (Fig. 7c,d). Under normal growing conditions, most of the bacterized plants were significantly increased in both root and shoot biomass compared to the non-bacterized plants. Thus, halotolerant biofilm-producing rhizobacteria effectively promote the growth of tomato plants under both non-stress and seawater-induced salt stress conditions.

Na + biosorption and ionic balance in plants.
To analyze how halotolerant biofilm-producing PGPR promoted plant growth and ameliorated seawater-induced salt stress damages, two representative bacterial isolates, namely ESK12 and ESM4 were inoculated in SOBG or SOBG containing 5% NaCl for biofilm formation. The biomass biofilm generated by these isolates was analyzed after centrifugation using FTIR. The peak positions of the biomass biofilm produced by ESK12 in SOBG containing 5% NaCl were shifted from 1046. 95 www.nature.com/scientificreports/ as compared to the biomass biofilm produced by this bacterium in SOBG only (Fig. 8a). Several peaks in the biomass biofilm created in SOBG supplemented with 5% NaCl by ESM4 were also shifted when compared to the biomass biofilm generated in the SOBG only (data not shown). One explanation is that Na + ions are absorbed by the various functional groups present in the bacterial biomass biofilm. We then examined the effects of biofilm-producing PGPR isolates in maintaining the ionic balance in the roots and shoots of tomato plants. The inoculation of PGPR isolates significantly affected the accumulation of Na + , K + , Ca 2+ , and Mg 2+ in the plant tissues (Fig. 8). In roots, all the bacterized non-stressed plants significantly lowered the Na + content and increased the accumulation of K + . The isolates ESK12 and ESM14 lowered the accumulation of Na + by 24% and 22% in roots, respectively (Fig. 8b). ESK17, ESM4, and ESM17 also effectively reduced the accumulation of Na + in the roots (Fig. 8c-e). We found that ESM24 lessened Na + accumulation by 43% in shoots, followed by ESM12 and ESM14 (Fig. 8g). ESM12 and ESK17 also increased K + accumulation by 35% and 18%, respectively compared with the non-inoculated control plant. Surprisingly, ESM24 increased K + accumulation by about 87%, while ESM4 to a lower extent. Importantly, bacterial isolates also maintained a higher K + /Na + ratio under non-stress and seawater-induced salt stress conditions in roots (Fig. 8d). It is worthy that all of the bacterized plants under seawater-induced salt stress conditions increased the Ca 2+ content by 15-35% and the Mg 2+ content on an average by 50% in roots compared with the non-inoculated plants. In shoots, bacterized plants also maintained the ionic balance by a lowered accumulation of Na + and increased accumulation of K + , Ca 2+ , and Mg 2+ . Moreover, bacterized plants maintained a higher K + /Na + ratio in both non-stress and seawater-induced salt stress conditions in shoots (Fig. 8i). As for shoots, the accumulation of K + , Figure 7. Effects of PGPR inoculation on the biomass accumulation and chlorophyll contents in tomato plants, grown under non-stress and seawater-induced salt stress conditions. Pot-grown tomato plants were inoculated with bacterial strains two times as described in "Methods" section. Analysis of (a) chlorophyll a content, (b) chlorophyll b content, (c) shoot dry weight, and (d) root dry weight. Blue and red bars in the graphs represent non-stress and seawater-induced salt stress. Values [mean ± standard error (SE)] of each treatment were attained from three biological replications (n = 3). Different letters represent significantly different data according to Fisher's least significant difference (LSD) (P < 0.05). FW fresh weight. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Ca 2+ , and Mg 2+ were also showed an increasing trend (Fig. 8h-k). In general, the bacterized plants diminished Na + accumulation and elevate K + accumulation in both roots and shoots. These results suggest that halotolerant biofilm-producing PGPR especially ESK12, ESK17, ESM12, and ESM14 efficiently lowered the Na + content and enhanced the K + , Ca 2+ and Mg 2+ content as well as the K + /Na + ratio in the roots. Thus, these bacteria effectively retain ion homeostasis, which are essential for maintaining plant growth and developmental processes under seawater-induced salinity.

Discussion
Forming a biofilm is one of the key salt-tolerant characteristics of rhizobacteria 39,40  Secondary metabolites [including osmoprotectants or compatible solutes (e.g., proline and glycine betaine), volatile organic compounds (e.g., albuterol, 1,3-propanediol and acetoin) and extracellular polymeric substances (EPS)] produced by various halotolerant PGPR play a crucial role in ameliorating salinity stress 1,40 . Among the secondary metabolites, EPS contain hydroxyl, carboxyl and amino functional groups 36,37 . These functional groups were reported to ameliorate salt-stress in plants through binding with cations including Na +39-41 . EPS also diminish the ion toxicity by minimizing the Na + influx through an increased expression of the HKT1/K + transporter in crop plants 66 . On the other hand, Dodd and Pérez-Alfocea 67 have stated that EPS generate a rhizosheath around the plant's roots, preventing the entry of Na + into plants, thereby reducing salt stress. The rhizosheath was also shown to increase soil aggregation, water-holding capacity, and acquisition of nutrients (N, P, K and Fe) from soil, leading to increased plant growth 68 . EPS-producing Halomonas sp. EX01 and Alcaligenes sp. were demonstrated to increase osmotic stress tolerance and stimulate rice growth under salt-stress condition 69,70 . Furthermore, EPS-producing G. halophytocola KLBMP and Pantoea alhagi were reported to activate the antioxidant system, helping to withstand salt stress damage in crops 71,72 . Several studies also reported that PGPR producing IAA, ACC deaminase, catalase and acetoin alleviate the salinity stress in plants 1,29,60,[73][74][75] . Our present study showed that the biofilm matrix and/or EPS contain several functional groups due to proteinacious curli, nanocelluloserich polysaccharides lipids and nucleic acids. Our FTIR data also revealed that biomass biofilm absorbed the Na + ions. Thus, the production of EPS, IAA, ACC, catalase and acetoin by PGPR might be one reason for alleviating the seawater-induced salt stress in tomatoes.
EL and RWC are two important parameters that determine the stress tolerance level in plants. Indeed, various abiotic stressors impair EL and RWC in plants. Among the abiotic stressors, salinity stress increases the EL 76,77 , but decreases the RWC 29,78-80 . A reduced RWC results in closing the stomata pore, diminishing photosynthesis, and accelerating senescence processes 81,82 . Moreover, depletion of RWC owing to salinity induces osmotic stress in plants 81 . Several studies have shown that PGPR-inoculated plants decreased the EL and increased the RWC in non-stress and under salt stress conditions 28,79,[83][84][85] . Osmoprotactants (e.g., glycine betaine, proline, proteins and sugars) synthesized by different halotolerant PGPR strains have been shown to prevent membrane damage and water loss under salinity stress 46,83 . In the present study, bacterized plants considerably reduced the EL, and maintained a higher RWC than non-bacterized plants in seawater-induced salt stress conditions. In this study, all the non-inoculated plants synthesized significantly more proline than the bacterized plants in both non-stress and seawater-induced salt stress conditions (data not shown). The results of the study indicated that non-inoculated plants experience higher stress than bacterized plants because proline is one of the important stress marker 86 . However, all the bacterized plant leaves (except when inoculated with ESM12 under seawater-induced salt stress conditions) exhibited a significantly higher total protein content than the non-bacterized plants in both nonstress and seawater-induced salt stress conditions. Plant cells accumulate ROS and degraded products of lipid peroxides [e.g. MDA, formaldehyde, acrolein, and 4-hydroxy-(E)-2-nonenal] which cause tissue damage during heat, drought, and salt stress 87,88 . To prevent the adverse effect of ROS, stressed plants induce an antioxidant defense system that scavenges the excess of ROS 7,18,23,87 . This study showed that seawater-treated non-inoculated plants accumulated higher levels of H 2 O 2 and MDA, while bacterized plants effectively suppressed these levels. Accordingly, %DPPH radical scavenging activity, CAT activity and total protein production were dramatically increased in bacterized plants under seawater-induced salt stress conditions. Among them, the tetrameric heme-containing enzyme, CAT (E.C.1.11.1.6) is catalyzing H 2 O 2 89 . Due to their antioxidant and iron binding properties, the total protein content plays a vital role in detoxifying ROS and free radicals, preventing lipid peroxidation 90 . The accumulation of anthocyanin in leaves improves plant adaptability to environmental stresses. Kitayama et al. 91 have stated that plants exposed to salt stress synthesized higher amounts of anthocyanins, suggesting that plants suffer oxidative stress. In the present study, less anthocyanin accumulation was observed in bacterized plants than in non-bacterized plants under seawater-induced salt stress conditions. Recently, a higher antioxidant property was reported in EPS produced by halotolerant PGPR 40 . EPS produced by the halophytic endophyte Glutamicibacter halophytocola KLBMP prevented the oxidative damage in crops triggered by salinity 71  www.nature.com/scientificreports/ stressed plants 46 . Indeed, PGPR producing ACC also enhanced salt tolerance and antioxidant enzyme activities, including CAT, in okra 75 . In this study, bacterized plants alleviated the seawater-induced salt stress in tomato by suppressing ROS levels and membrane lipid peroxidation. Bacterized plants also reduced the seawater-induced salt stress through a higher %DPPH radical scavenging-and CAT activity, total protein content, and anthocyanin synthesis. These mechanisms enable bacterized plants to promote chlorophyll synthesis and subsequently biomass accumulation. In the current study, ESK12, ESM4, ESM12 and ESM14 inoculation resulted in an increased biomass in both roots and shoots by 30-50% as compared to non-bacterized control plants under saline conditions. All the inoculated bacterial isolates efficiently colonized the roots in both non-stress (3.4E+9 to 4.6E+9 CFU g −1 ) and seawater-induced salt stress (3.6E+9 to 4.9E+9 CFU g −1 ) conditions (data not shown). Increased photosynthesis due to higher chlorophyll content results in higher photosynthates and finally in higher biomass accumulation. In this study, ESK12, ESM12, ESM17, and ESM24 significantly increased the chl a content and ESK12 and ESM14 the chl b content in tomato plants. Similar reports have shown that salt tolerant PGPR including Achromobacter piechaudii and P. stutzeri improved salt tolerance and increased the biomass accumulation of tomato seedlings under high salinity stress 92,93 . However, EPS enclosed with biofilms were not studied earlier to mitigate seawater-induced salt stress in tomatoes.
In response to salt stress, plants adjust ionic balance and cellular metabolism to continue growth and developmental processes 94,95 . In general, Na + and Cl − disturb ionic balance within the cells and adversely affect plant growth due to oxidative injury caused by ROS and lipid peroxidation 88 . The accumulation of Na + within cells inhibits the absorption of K + and translocation to the cellular compartment 44 . K + is an essential element of various physiological processes in plants. It accelerates water movement and enzyme activation, which finally affects adenosine triphosphate (ATP) production and the rate of photosynthesis 96 . Na + accumulation in soil solutions also prevents absorption of macronutrients, such as Ca 2+ and Mg 2+ , which are essential for membrane stability and stomatal conductance 44 . Thus, increased accumulation of Na + is toxic, but increasing accumulation of K + , Ca 2+ and Mg 2+ are beneficial for the various metabolic processes 96 . This study showed that seawater-induced salt stress provoked oxidative stress in plants and eventually suppressed plant growth. However, bacterized plants significantly exhibit a lowered accumulation of Na + and ultimately maintained a higher K + influx in shoot under seawater-induced salt stress conditions. Notably, bacterized tomato plants also maintained a higher K + /Na + ratio in the roots and shoots. The approximation of Ca 2+ and Mg 2+ content in root and shoot biomass also showed that halotolerant biofilm-producing rhizobacteria increased their accumulation in both non-stress and seawaterinduced salt stress plants. Thus, halotolerant biofilm-producing rhizobacteria maintain the ionic balance and ion homeostasis by decreasing the accumulation of harmful Na + and increasing the accumulation of K + , Ca 2+ and Mg 2+ . In nature, 99% of bacteria form biofilms on different biotic and abiotic surfaces. In this study, Na + is bound by the biomass biofilm. Thus, it is believed that Na + bound by the bacterial biofilm formed around the roots by various halotolerant biofilm-producing bacteria, prevents Na + to enter into the cellular compartment. On the other hand, the absorption of K + and macro-elements including Ca 2+ and Mg 2+ are instead increased by the bacterization of tomatoes.
This study extends our understanding that halotolerant biofilm-producing rhizobacteria can enhance crop productivity under saline agro-ecosystems. The experimental results showed that the identified halotolerant biofilm-producing bacteria attenuated seawater-induced salt stress and enhanced biomass accumulation in tomato plants through the formation of biofilms, availability of nutrients (e.g. N, P, Fe, and K), production of IAA, ACC deaminase and NH 3 , expression of antioxidant defense systems, and maintenance of a proper ionic balance. Unfortunately, pure culture isolates with multiple PGP traits does not always enhance plant growth under field conditions, due to a competition with indigenous microflora and other environmental stresses. Other challenges that affect survival and colonization in roots under field conditions by PGPR are inappropriate carrier used for bio-formulation, lower numbers of CFU, lower shelf-life of the inoculants, and poor soil health owing to soil pollution and long-term fumigation. Still, we believe that halotolerant biofilm-producing bacteria with multiple PGP activities would be a beneficial approach in addressing the above challenges.

Isolation of halotolerant biofilm-producing rhizobacteria. Tomato plants (Solanum lycopersicum
L.) grown in the saline (3.5-5.0 dS m −1 ) affected areas, such as Patuakhali and Bagerhat district, Bangladesh were carefully uprooted during the vegetative stage. The rhizosphere samples were prepared as described by Haque et al. 64 . Each rhizosphere sample was then homogenized with sterile 0.5 × phosphate buffer saline (PBS) using a sterile mortar and pestle. Halotolerant rhizobacteria were isolated as explained by Jha et al. 97 with minor modifications. In brief, 50 µL diluted sample was inoculated in glass test tubes containing 5 mL yeast extract peptone [YEP peptone (1%), yeast extract (0.5%), pH 6.8)] broth containing 5% NaCl instead of nitrogen-free semisolid NFb media containing malate and 4% NaCl. The inoculated test tubes were kept at 28 °C in a shaking condition (160 rpm). After a 72-h incubation, 100 µL suspension was transferred onto YEP agar (1.5%) plates (3 plates sample −1 ) supplemented with 5% NaCl. Colonies formed on the plates after 48 h incubation at 28 °C were herein termed as halotolerant rhizobacteria. Morphologically diverse (e.g., size, shape and pigmentation) and well separated from others, five colonies (isolates) from each sample were picked using sterile toothpicks, and sub-cultured to prepare a pure culture. All the halotolerant rhizobacterial isolates were screened for biofilm formation as described by Haque et al. 98 . In brief, each isolate was initially cultured in YEP broth overnight. Then, 50 μL culture (ca. 10 6 CFU mL −1 ) was injected the flat bottom glass test tubes (Pyrex, UK) containing saltoptimized broth [SOB (for 1 L: 5 g yeast extract, 20 g tryptone, 2.4 g magnesium sulphate, 0.5 g sodium chloride, and 0.186 g potassium chloride)] containing 2% glycerol (SOBG) media. The test tubes were set at 28 °C without shaking. The AL (air-liquid) and SAL (solid-air-liquid) biofilms were categorized and pictures were taken. To quantify the formation of biomass biofilm, 72-h old biofilm matrices were carefully transferred into 15 mL falcon www.nature.com/scientificreports/ tubes and centrifuged (15,000 rpm for 10 min). After centrifugation, each pellet (biofilm matrix) was weighed using a digital balance. The CFU from biofilm matrices was quantified as explained 99 .

Identification of bacteria.
Bacterial genomic DNAs were isolated using a standard protocol 100  Identification of the biofilm matrix components. In order to identify the biomolecules and chemical functional groups/ligands present in the biofilm matrices/EPS, the FTIR spectroscopy was used. Initially, biofilm matrices were harvested by centrifugation (12,000 rpm for 10 min) to discard the liquid. The pellets were further analyzed. To identify the components, 450-4000 cm −1 was scanned (16 scans at 4 cm −1 resolution and at 0.2 cm s −1 scanning speed) using the TGS (Triglycine sulphate) detector. The IR spectra of the biofilm matrices were acquired using the Perkin Elmer FTIR (Spectrum-2) instrument and the CPU32M software. To analyze the baseline subtracted biofilm spectra, Perkin Elmer's proprietary software (Version 10.05.03) was used. Curli and/ or cellulose production were checked by Congo red-and Calcofluor binding assays 99 . In this case, 2 μL culture (overnight grown) was dotted onto the agar plates supplemented either Congo red (40 μg mL −1 ) or Calcofluor white (200 μg mL −1 ). The plates were then kept at 28 °C for 48 h. Calcofluor binding (but not the Congo red binding) was monitored under UV light (366 nm) and photos taken. Nanocellulose fibers present in the bacterial biofilm matrices were observed by SEM (JEOL JSM-6490LA, Japan). For SEM analysis, 72-h old biofilm matrices were carefully collected and centrifuged at 10,000 rpm for 10 min. The pellets were gently oven-dried (50 °C for 24 h) 33 . Each pellet was fixed on adhesive carbon tape and coated with platinum using a vacuum coater. The images were taken by SEM operated at 5.0 kV.

Preparation of inoculum for plant growth-promoting tests. A single colony of each isolate was
injected in the glass test tube containing YEP broth and incubated at 28 °C in agitated condition (160 rpm) until optical density (OD 660 ) reached 0.6-0.8. Each culture (1 mL) was then centrifuged at 12,000 rpm for 10 min. Each pellet was re-suspended in YEP broth. In order to bioassays of the in vitro PGP activities, 2 µL (ca. 10 7 CFU mL −1 ) culture was spotted onto agar plates (4 dots plate −1 ), while 50 µL culture was inoculated in YEP broth.
In vitro plant growth promoting tests. All the PGP tests were performed using the respective medium (as mentioned below) supplemented with 5% NaCl. IAA production was quantified using the Salkowsky colorimetric-and HPLC method as described earlier by Gordon and Weber 101 and Myo et al. 102 , respectively. In both the cases, all the biofilm-producing isolates were grown in LB (Luria Bertani) broth supplemented with L-tryptophan (0.2%) in agitated condition. For colorimetric determination, Salkowsky reagent (35% perchloric acid and 0.5 M FeCl 3 ) was mixed with each culture supernatant and kept at room temperature in dark. After 45 min, IAA (an auxin) was estimated by spectrophotometer with an absorbance at 530 nm. Different known concentrations (0-1000 μg mL −1 ) of IAA (Sigma-Aldrich, St. Louis, MO, USA) were used as standard. For HPLC analysis, cell-free extract was acidified then the organic phase was collected. After evaporation of ethyl acetate, HPLC-grade methanol was added to dissolve the crude extract. Shimadzu Prominence HPLC (Japan) with a C18 analytical column (4.6 mm × 250 mm, 5 μm) were used to quantify IAA. Methanol and 1% acetic acid (50:50 v/v) were used as the mobile phase. Column temperature, flow rate and injection volume were maintained at 25 °C, 1 mL min -1 and 20 μL, respectively. Lab Solutions software was used to evaluate the data, comparing it with the elution profiles of standard IAA. Nitrogen fixation study was done according to Ker 103 . In addition, two forward primers, F1-TAYGGIAARG-GIGGIATYGGIAARTC and F2-TGY GAY CCIAAIGCIGA, and one reverse primer, R6-TCIGGI GAR ATG ATGGC were used to amplify the 470 bp and 359 bp nifH fragment by PCR 104 64 . The P solubilization index (PSI) was also calculated 64 . For K solubilization, 2 µL culture of each isolate was dotted onto Aleksandrov agar plates [for 1 L: glucose 5.0 g, MgSO 4 ·7H 2 O 0.5 g, FeCl 3 0.005 g, CaCO 3 0.1 g, Ca 3 (PO 4 ) 2 2 g, AlKO 6 Si 2 (potassium aluminum silicate) 2 g, and agar 15 g] 106 . AlKO 6 Si 2 was used as a source of insoluble inorganic potassium. For Zn solubilization, 2 µL suspension of each bacterium was dotted onto basal agar plates (per 1 L: glucose 10 g, (NH 4 ) 2 SO 4 1 g, KCl 0.2 g, K 2 HPO 4 0.1 g, MgSO 4 ·7H 2 O, and agar 15 g, pH adjusted to 7.0) containing 0.2% insoluble Zn 106 . In this study, Zn 3 (PO 4 ) 2, ZnCO 3 , and ZnO (Wako Pure Chemical Industries Ltd., Japan) were used as insoluble Zn sources. All the spotted plates (P, K, or Zn solubilization) were placed at 28 °C. Development of a clear zone surrounding the colonies within 7 days suggested a positive P, K, or Zn solubilization. In order to detect the production of siderophore i.e., Fe solubilization, 2 µL culture of each isolate was dotted onto the LB agar plate and incubated. After  www.nature.com/scientificreports/ broth 107 was applied over those LB agar plates. The color change from blue to light yellow, yellow, light orange/ orange or purple indicated a positive siderophore production. Based on color, the siderophore was classified as catechol (purple), hydroxamate (light orange/orange), carboxylate (light yellow), and mixture of hydroxamateand carboxylate (yellow) 57 . Qualitative production of volatile components including acetoin, indole, NH 3 and HCN were determined as described by Dye 108 , Lelliott and Dickey 109 , Dinesh et al. 110 , and Lorck 111 , respectively. For acetoin production, 30 μL bacterial culture was mixed with 3 mL broth of yeast extract salt and incubated under shaking condition. After 72 h, 500 μL culture was mixed with 300 μL α-napthol (5%) in absolute alcohol (v/v). Formation of a crimson to ruby color on the surface within 5 h indicated a positive acetoin production. For indole production, 30 μL bacterial suspension was injected in 3 mL broth containing tryptone (10 g L −1 ), l-tryptophan (1 g L −1 ) and the required amount of water (L −1 ) and the tubes were placed at 28 °C in 160 rpm. After 3 days, Kovac's reagent (250 μL) was added and mixed gently. Formation of a dark red color on the surface of the mixture indicated a positive indole production. For NH 3 production, 30 μL culture was suspended in the tube containing 2.5 mL peptone water and kept at 28 °C with 160 rpm. After 3 days, Nessler's reagent (500 μL) was added to each test tube. Yellow to brown color formation confirmed a positive NH 3 production. In case of HCN production, a single colony of each bacterium was streaked on YEP plates supplemented with glycine (0.45%). Then, three sterilized filter papers were soaked in picric acid (0.25%) and sodium carbonate (1.25%) solution and placed on the lids of petri plates. The plates were then sealed and kept at 28 °C for 24 h. The color change of the filter papers from greenish to yellow, reddish or brown indicated a positive HCN production.
The rhizobacterial isolates were screened for ACC deaminase activity on the sterile minimal DF (Dworkin and Foster) salts media 112 supplemented with 3 mM ACC (1-cyclopropoane-1-carboxylic acid) as the sole nitrogen source 113 . In brief, each isolate (50 μL) was plated on minimal Dworkin and Foster (DF) salt agar plates. A 50 μL suspension was also plated onto ACC-deficient minimal DF salt agar plates as a negative control. The plates were incubated at 28 °C for 5 days. Colonies formed on the plates were considered as ACC-deaminase producers. Catalase, oxidase, gelatinase, and arginine dihydrolase were also qualitatively determined as described in Hayward 114 ,Shekhawat et al. 115 ,Schaad 116 , and Thornley 117 , respectively, while lipase 110 , cellulase, and protease assays were carried out as described 118,119 . For catalase production, a loop-full (24-h old) culture was placed on a clean glass slide then a drop of 3% H 2 O 2 was added. Production of gas bubbles indicated a positive reaction. For oxidase test, a strip of filter paper was soaked with 3 drops of N,N,Nˊ,Nˊ-Tetramethyl-p-phenylenediamine dihydrochloride (color indicator). Then, a loop-full overnight-grown culture of each isolate was smeared onto the surface of the filter paper. Development of a purple color within 10 s indicated a positive reaction. For gelatinase production, a medium (1 L) containing beef extract (3 g), gelatin (120 g), peptone (5 g) and an adequate amount of water was prepared. All ingredients were dissolved by steam heating, then 3 mL suspension was poured in test tubes and autoclaved. After cooling the tube, stab inoculation was done for each bacterium and the tubes placed at 28 °C. After 72 h, the test tubes were transferred to 4 °C for 30 min. A flow of the medium after gently tipping indicated a positive gelatinase production. In case of lipase production, 1 μL suspension of each isolate was dotted onto YEP agar plates (4 dots plate −1 ) containing Tween 20 (1%). Development of translucent zones around colonies within 48 h incubation at 28 °C confirmed a positive lipase production. For cellulase production, 1 µL suspension of each isolate was marked onto LB agar plates supplemented with carboxymethyl cellulose (1%) and placed at 28 °C. After 24 h, the plates were stained with 0.01% Congo red for 15 min, followed by de-staining with 1 M NaCl. Formation of yellowish zones around bacterial colonies indicated positive results. In case of protease production, 1 µL suspension was dotted onto LB agar plates containing 1% skim milk and incubated at 28°C. Formation of clear zones around colonies within 3 days suggested a positive protease production. In all the cases, Pseudomonas chlororaphis ESR15 64 was used as a positive control, while Bacillus cereus ESD3 43 was only used as a positive control for the protease test. All these experiments were repeated three times with 4 replications.
Abiotic stress tolerance tests. All the isolates were confirmed for their ability to grow on agar plates at different concentrations of NaCl (15, 10 and 5%), temperatures (50, 42 and 37 °C) and varying pHs (10, 9, 8, 7 and 4). In all these assays, a 50-µL aliquot was plated on LB agar plates. In order to examine the effect of NaCl and pH, the plates were incubated at 28 °C. In contrast, the plates were incubated at the desired temperatures to study the effect of different temperatures on growth. In all cases, the growth was examined after 96 h of incubation. B. cereus ESD3 was used as a positive control 43 . All tests were performed twice with three replications.
In vitro antagonistic activities. All the halotolerant isolates were further tested for their antagonistic activities against X. campestris pv. campestris ATCC 33913 (causal agent of bacterial leaf spot in tomato), R. solanacearum ATCC® 11696™ (causal agent of wilt in tomato), and P. carotovorum subsp. carotovorum PCC8 (causal agent of soft rot in tomato, accession number KX098362) as described by Furuya et al. 120 . In brief, each rhizobacterial isolate (1 μL) was dotted onto YEP agar plate and kept for 48 h. A sterilized filter paper soaked in chloroform was placed in the lid of the Petri dish to kill the rhizobacterium. The petri dishes were placed at room temperature for 1.5 h. Then, melted water agar (50 °C) containing ca. 10 9 CFU mL -1 plant pathogenic bacterium was decanted onto the plates and incubated at 28 °C. Development of an inhibition zone nearby the growth region of the rhizobacterium within 48 h indicated a positive biocontrol agent.
Setting of pot experiment. BARI (Bangladesh Agricultural Research Institute) Tomato 2 was used in this study. Seedlings were grown according to the method described in Haque et al. 64 . In brief, healthy seeds were surface-sterilized using 5% NaOCl followed by 4 times washing with sterile distilled water. Then, the seeds were sown in a plastic tray containing sterile sand. For pot experiment, 3 parts of sterilized vermicompost were mixed with 1 part of sterilized garden soil (1.7% organic matter, 0.09% total N, 1.75 mg kg −

Electrolyte leakage.
Leaf discs of the third leaf from the top were used to determine electrolyte leakage (EL). Briefly, 1 cm 2 of 5 leaf discs were transferred into glass tubes containing 10 mL distilled water. The electrical conductivity (EC) of the leaf sample was measured at room temperature twice after incubation in water bath at 40 °C and 100 °C for 30 min, respectively 121 . Then, the EL was calculated using the following formula: Relative water content. Five leaf discs (1 cm 2 ) from the top third leaf were taken and immediately the fresh weight was measured to determine the relative water content (RWC). The turgid weight of the leaf discs was measured after incubation in distilled water at 4 °C for 24 h. Then the leaf discs were oven-dried at 72 °C for 48 h and the dried weight was recorded. The RWC was calculated as the following formula: Quantification of hydrogen peroxide levels. Hydrogen peroxide (H 2 O 2 ) concentration was determined according to Loreto and Velikova 122 . For this, 0.5 g leaf samples were homogenized in 3 ml 1% (w/v) tri-chloroacetic acid (TCA). The homogenate was centrifuged at 10,000 rpm for 10 min at 4 °C. Subsequently, 0.75 mL of the supernatant was added to 0.75 mL of 10 mM phosphate buffer (pH 7) and 1. Quantification of lipid peroxidation. The lipid peroxidation level was quantified by the thiobarbituric acid (TBA) reaction as described by Vemanna et al. 123 . The representative lipid peroxidation product malondialdehyde (MDA) was quantified from bacterized and non-bacterized leaf samples of plants from both non-stress and seawater-induced salt stress conditions. Fresh leaf samples (0.2 g) were ground in 2.5 mL of 0.1% trichloroacetic acid (TCA) and centrifuged at 14,000 rpm for 15 min. After centrifugation, 1 mL of the supernatant was mixed with 2.5 mL 0.5% TBA in 20% TCA and incubated in a hot water bath (95 °C) for 30 min. Thereafter, the reaction was stopped immediately on ice and then centrifuged at 10,000 rpm for 10 min. Absorbance at 532 nm and 600 nm was determined, and MDA concentration was estimated by subtracting the non-specific absorption at 600 nm from the absorption at 532 nm using an absorbance coefficient of extinction (155 mM −1 cm −1 ).

Measurement of leaf anthocyanin contents.
Leaf anthocyanin contents were spectrophotometrically estimated by differential pH method. For this, leaf samples (0.2 mg) were ground with 2 mL extraction buffer www.nature.com/scientificreports/ acetate buffer (pH 4.5, 0.4 M). Then the absorbance was recorded at 510 nm and 700 nm, respectively and water was used as blank 124 . Mean absorbance (A) of the two buffers was calculated as follows: Total anthocyanin content was estimated using the following equation: where A is the absorbance of the sample, DF is the dilution factor (10), MW is the molecular weight of cyaniding-3-glucoside (449.2), a major anthocyanin, ε is the molar extinction coefficient of cyaniding-3-glucoside (26,900 in L × mol -1 cm -1 ) and 1 is the path length (1 cm).
Quantification of catalase activity. Total protein was first extracted from the leaf sample to measure the catalase (CAT) activity. About 0.2 mg leaf sample was homogenized in phosphate buffer (50 mM, pH 7.0) containing potassium chloride (100 mM), ascorbate (1 mM), β-mercaptoethanol (5 mM) and glycerol (10%, v/v). Then, the homogenized sample was centrifuged at 14,000 rpm for 15 min at 4 °C. One mL reaction mixture was prepared for phosphate buffer (100 mM, pH 7.2), EDTA (0.1 µM), and H 2 O 2 (0.1%). The reaction was started after adding 20 µg of protein and then decrease of H 2 O 2 was monitored at 240 nm with a spectrophotometer (APEL PD-303, Japan). CAT activity was expressed as µmol min −1 mg −1 protein.
%DPPH radical scavenging activity. The total antioxidant activity of the leaf samples was estimated by the radical scavenging activity using the 1,1-diphenyl-2-picryl hydrazyl (DPPH) method 125 . Briefly, 0.2 g leaf was crushed in phosphate buffer (100 mM, pH 7.0). After centrifugation at 14,000 rpm for 15 min at 4 °C, 0.1 mL supernatant was added to the DPPH solution (0.9 mL of 0.2 mM DPPH in ethanol). Then, the reaction mixture was incubated at 25 °C for 30 min in dark, after which the absorbance was measured at 517 nm. Instead of plant extract, the equal amount of methanol was added to DPPH solution as a control. The percent DPPH radical scavenging activity was estimated as follows: where A 0 is the absorbance of the control and A 1 is the absorbance of the plant extracts.
Measurement of chlorophyll content. Chlorophyll (Chl) contents were extracted from 0.2 g of leaf tissues in 5 mL of acetone (90%) and then kept at room temperature for 48-h in the dark. The supernatant was topped to a final volume of 10 mL. The absorbance was recorded at 663 nm and 645 nm using a UV-visible spectrophotometer. Chl a and Chl b were estimated by using the following formula and expressed as mg g −1 FW 51 .
Estimation of biomass accumulation. Forty-day-old plants were harvested from the pots to quantify biomass accumulation in the roots as well as in the shoots. The root-adhering soils were carefully removed by gentle washing with running tap water. Then the roots and shoots were separated and kept in brown paper for oven drying at 70 °C until a constant weight was achieved. Dry weight was recorded to calculate the biomass accumulation in plants.

Statistical analysis.
A one-way analysis of variance (ANOVA) test was done for the in vitro experiments, while multiple factor ANOVA test was performed for the in vivo pot experiments. In all the cases, Shapiro-Wilk test was used to confirm the normal distribution of the data sets. ANOVA, distribution of data, homogeneity of variance, and mean comparison of treatment effects were analyzed using the R software version 3.3.6. Fisher's least significant difference test was applied to compare the means. www.nature.com/scientificreports/