The bacterial strains and the plasmid constructs used in this study were listed in Table 1. P. phytofirmans PsJN and its derivatives were cultured in KB (King’s B medium), LB (Luria Broth), ½ (half strength) PDA/PDB (Potato Dextrose Agar/Broth, Difco™ USA), NA (Nutrient Agar), R2A or TSA (Tryptone Soya Agar) at 28°C.

The in-frame deletion of bceQ in the genome of PsJN was performed by following a previous method (Yan et al., 2017). Briefly, a deletion construct pEX18TC-ΔbceQ which contains 500 bp of upstream and 506 bp of downstream sequences flanking the bceQ gene, was amplified by PCR using primers bceQ-up-F and bceQ-up-R, bceQ-down-F, and bceQ-down-R (Table 1), respectively. The deletion construct was transformed into wild-type PsJN and the transformants were cultured on KB plates amended with tetracycline at 20 μg/ml to select colonies that had the deletion construct integrated into the bacterial genome. The transformants were then cultured on KB plates plus 5% sucrose to identify colonies that have the deletion mutation by PCR analysis. Specifically, the wild type PsJN produced a 1,540 bp-DNA fragment and the ΔbceQ mutant produced a 163 bp-DNA fragment, respectively (Supplementary Figure S1), using the primers bceQ-Full-F and bceQ-Full-R (Table 1).

The complementary strain ΔbceQ-Comp (LK751) was generated by introducing the construct pME6010-bceQfull into ΔbceQ mutant. pME6010-bceQfull was made by cloning a 1,540 bp-DNA fragment which contains the bceQ full gene into pME6010 at the Xhol and HindIII restriction enzyme sites. The complementary strain ΔbceQ-Comp was confirmed by PCR analysis (Supplementary Figure S1). Strains LK749 and LK750 were made by introducing the empty vector pME6010 into the wild type PsJN and the ΔbceQ mutant, respectively.

PsJN and its derivatives were cultured in ½ PDB liquid medium at 28°C with shaking for 3 days. The cultures were centrifugated for 5 min at 8,000 r.p.m. to remove the cells and collect the supernatant. To extract the EPS, pure ethanol proof 200 was mixed with the culture supernatant at a ratio of 2:1 (ethanol: supernatant, v/v). The EPS was precipitated by incubating the mixtures at −20°C overnight and centrifugated for 10 min at 12,000 r.p.m. The clear supernatants were carefully discarded, and the pelleted EPS was dried at 45°C overnight, weighed, and the dry weight of EPS was normalized by the volume of the bacterial cultures. The experiment was repeated three times independently.

Bacterial motility is important in the interaction with the plant host and can be influenced by EPS production. The swimming and swarming motility assays were conducted on 0.2% agar and 0.6% agar of TSA plates, respectively. Specifically, 2 μl of the bacterial cell suspension (OD₆₀₀ = 0.1, equals around 10⁸ cfu/ml) were inoculated on the fresh media plates. The plates were incubated at 28°C without disturbance for 3 days before the diameters (cm) of the spreading colonies were measured. The experiment was repeated two times independently with three replicates of all the tested strains in each experiment.

PsJN may encounter salt and/or iron stresses in environments because of the accumulation of different salts via natural processes and agricultural practices (Kumar et al., 2020). To test the bacterial resistance to salt and iron stresses, NaCl and FeCl₃ were added to the KB liquid medium at a final concentration of 100 and 1 mM, respectively, based on our preliminary experiments. The growth assays were conducted in 96-well plates. Specifically, the bacterial cells of PsJN and the ΔbceQ mutant were inoculated in KB broth (with and without salt amendments) at a starting OD₆₀₀ = 0.001 (around 10⁶ cfu/ml). A 250 μl of the starting cultures was aliquoted into a 96-well plate with each strain having four replications. The 96-well plate was incubated in a SPARK® multimode Microplate Reader (TECAN, Switzerland) with continuous shaking (around 200 r.p.m.) at room temperature. The cell density was measured every 2 h by recording the absorbance of optical density at 600 nm. All the growth assay experiments were conducted two times independently, with each experiment including four replicates of the tested strains.

Wild type PsJN and the ΔbceQ mutant were inoculated on sterilized filter papers (Millipore, 0.45 μm, Bedford Massachusetts) that were placed on ½ PDA plates. The plates were incubated at 28°C for 3 days before the following desiccation or UV stress assays. For the desiccation resistance assay, the bacterial colonies, carried by the filter papers, were removed from the culture plates and dried in a sterilized Petri dish at 28°C. After drying, the population size of the bacterial colonies (cfu/colony) was measured immediately (day 0), and on 2, 8, and 14 days. Specifically, the filter paper carrying bacterial colonies was cut into small pieces with one colony per piece of the filter paper. The population size of the colonies was measured by placing one piece of filter paper in 1 ml of sterilized water followed by serial dilution plating on PDA agar plates. For the UV stress resistance assay, the bacterial colonies were removed from the culture plates and placed in a sterilized Petri dish as described above. The colonies were directly exposed to UV radiation in a lumina flow hood (Purifier Logic+ Class II, Type A2 Biosafety Cabinet, Labconco Corporation). The bacterial colonies were sampled at 0, 10, 30, and 50 min after the UV exposure and the population size (cfu/colony) was measured by dilution and plating assay as described above. Both the desiccation and the UV resistance experiments were repeated three times independently with each experiment having four replicates of the tested strains at each time point.

The bacteriophage tolerance assay was conducted by following the previous report with modifications (Maynard et al., 2010). The bacteriophage vB_Pap-BS21 (Paraburkholderia phytofirmans strain PsJN, Bozeman soil, 2021) was isolated from the greenhouse pot soils following the previous report (Fu et al., 2021). The phage was purified through five rounds of plaque purification on bacterial lawn plates. To test the impact of the isolated bacteriophage on the growth of PsJN, the same volume of the bacterial cell suspension (OD₆₀₀ = 0.1, equals around10⁸ cfu/ml) was mixed with the phage particle suspension (10⁶ pfu/ml) to reach a ratio of phage particle: bacterial cell equals 1:100 in ½ PDB liquid medium. The mixtures were aliquoted into a 96-well plate with 250 μl per well. The 96-well plate was incubated in the Tecan Microplate Reader with continuous shaking at room temperature and the cell density was recorded by measuring the absorbance of optical density at 600 nm. The experiments were repeated two times independently, with each experiment included eight replicates of the tested strains.

TEM was used to examine bacterial cell morphology. Bacterial cells that were collected from fresh cultures on KB plates were negatively stained using the negative stain solution 1% phosphotungstic acid before the TEM analysis. The experiment was conducted at the TEM Center of Montana State University.

Camelina (Camelina sativa cv. Suneson) and common pea (Pisum sativum) supported a decent endophytic colonization by PsJN in our preliminary experiments and were used in this study. The seeds were sowed in 6-inch pots containing soils (pH: 7.3; NO₃-N: 68 ppm; PO₄-P: 3 ppm; potassium: 47 ppm) that were autoclaved twice at 121°C for 2 h. The pots were kept in a greenhouse that was maintained at 22/18 ± 1°C (day/night), a humidity level of 42%, and a photoperiod of 16/8 h light/darkness, at the Plant Growth Center, Bozeman, Montana. Regular watering was provided based on the moisture level of the soil. Drought stress, if needed, was introduced by stopping watering the plants immediately after the bacterial inoculation (as described below). The plant samples were collected and analyzed on day 7 after inoculation when the drought-induced wilting symptoms were visibly shown.

Leaves of four-week-old camelina and three-week-old pea were inoculated with the bacterial strains and tested for their plant colonization. Specifically, the bacterial suspension (OD₆₀₀ = 0.1) was injected into leaves using a needleless syringe until the inoculation covered most areas of the leaf. Each bacterial strain was tested on three plants, with each plant having three leaves treated. The bacterial colonization was enumerated on days 0 and 7 by serial dilution plating of the extracts of a leaf disc (around 1 cm²) onto KB plates added with Rifampicin (100 μg/ml). The plates were incubated at 28°C until colonies developed, which typically took 3 days.

Statistical analysis was performed using the Samples t-test function of the SPSS software. A p-value of less than 0.05 was considered a significant difference. The sample sizes and number of replicates of each specific experiment were indicated in the respective methods described above.

PsJN can grow on different media although the biosynthesis of EPS might be favorited under specific culture conditions. To identify the culture media that favorite the EPS production, the wild type PsJN was cultured on agar plates that were made of seven different culture media, including NA, NA + Glycerol, LB, KB, half strength (½) PDA, R2A, and TSA. As shown in Figure 1A, PsJN formed more mucoid colonies, indicating a higher level of EPS production, on NA + Glycerol and ½ PDA plates than the other tested media. Further, PsJN grew slightly faster on ½ PDA than on NA + Glycerol (data not shown). Thus, the ½ PDA medium was used in the following EPS assay.

The genome of PsJN contains a cluster of 19 genes that encode proteins that are putatively responsible for the biosynthesis, translocation, and export of EPS (Figure 1B). To characterize the role of the EPS biosynthesis genes in EPS production, an in-frame deletion of bceQ, encoding a putative flippase for the translocation of EPS intermediates from inner membrane to periplasmic spaces, was introduced in the genome of wild type PsJN (Figure 1B; Supplementary Figure S1). The resultant ΔbceQ mutant formed much less mucoid colonies than the wild type PsJN on culture plates (Figure 1C). The complementary strain ΔbceQ-Comp, containing the wild type bceQ gene on a plasmid, restored the colony morphology similar to the wild type strain. To quantify the EPS production, PsJN and its derivatives were cultured in ½ PDB liquid medium, and the produced EPS was extracted and measured. As shown in Figure 1D, the ΔbceQ mutant produced EPS at a significantly lower level than the wild type and the complementary strain (Figure 1D), thus validating the role of bceQ in the EPS production of PsJN. Interestingly, a dark layer of the extracellular substrate was observed surrounding cells of the wild type PsJN, the complementary strain, but not the ΔbceQ mutant (Figures 1E–G), which is consistent with the known function of EPS as the extracellular polysaccharide on the outside surfaces of the producing cells.

PsJN and its derivatives were cultured on TSA plates that contained 0.2 and 0.6% of agar to assess the swimming motility and swarming motility, respectively. The results showed that the ΔbceQ mutant had a significantly decreased swimming motility, as determined by the decreased diameter size of the bacterial colony spreading on the culture plates, than the wild type and the complementary strains (Figure 2). No significant difference of the swarming motility was observed between the wild type PsJN and the ΔbceQ mutant (data not shown). This result suggests that EPS production positively influenced the swimming motility of PsJN. This result is consistent with the previous reports that EPS, similar to other polysaccharides, may serve as motility substrates facilitating bacterial movement (Gygi et al., 1995; Berleman et al., 2016).

To investigate if lack of EPS production affects PsJN’s growth under normal culture conditions or with the presence of salt and iron stress, the wild type PsJN and the ΔbceQ mutant were cultured in KB broth with and without the addition of NaCl and FeCl₃. As shown in Figure 3A, the ΔbceQ mutant had a similar growth curve as the wild type PsJN, indicating that lack of EPS production does not affect the growth of PsJN in the KB medium.

Adding NaCl at 100 mM did not influence the wild type PsJN’s growth but negatively affected the ΔbceQ mutant with a prolonged exponential growth stage (Figure 3B), indicating a decreased growth rate of the mutant under the salt stress. The addition of FeCl₃ at 1 mM decreased the growth of the wild type PsJN and the ΔbceQ mutant to a maximal OD6₀₀ of 0.8 (Figure 3C), which is lower than their maximal OD6₀₀ of 1.2 in KB without treatment. Importantly, the ΔbceQ mutant showed a clearly prolonged exponential growth profile, compared to the wild type PsJN, with the FeCl₃ treatment (Figure 3B).

Overall, these results suggested that EPS production was required for the maximal growth of PsJN under the salt and iron stresses.

To test if EPS protects PsJN from desiccation and UV damage, the wild type PsJN and the ΔbceQ mutant were inoculated on filter papers that were placed on the top of ½ PDA plates. The colonies were removed on day 3 and dried in an empty petri dish to test the bacterial survival under desiccation. As shown in Figure 4A, the populations of both wild type PsJN and ΔbceQ mutant decreased with the extended desiccation. However, the ΔbceQ mutant’s population declined faster than the wild type. For example, around 1% of the wild type PsJN’s population remained alive on day 14, but only less than 0.1% of the ΔbceQ mutant’s population survived at this time point.

To test the bacterial survival under UV stress, the above cultured fresh bacterial colonies were placed under a UV light, and the living cells were measured at different time points. The results showed that no living cells of the ΔbceQ mutant were detected at 30 min after UV exposure, but around 0.01% of the wild type’s population remained alive (Figure 4B).

These results suggested that EPS production was required by PsJN to tolerate desiccation and UV damage.

Bacteriophages are bacterial natural enemies that are ubiquitous in the environment and can affect bacterial survival and activity. In this work, a bacteriophage vB_Pap-BS21 isolated from greenhouse pot soils was found to infect PsJN and caused transparent plaques on culture plates (Figure 5, inner photo). To test if EPS plays a role in protecting PsJN cells against phage infection, the purified bacteriophage particles were inoculated in the bacterial cultures of the wild type PsJN and the ΔbceQ mutant. Inoculation of the bacteriophage caused a clear growth deficiency at the early exponential stage of the ΔbceQ mutant, compared to the wild type PsJN and the complementary strain (Figure 5). All the tested strains eventually reached the stationary phase, suggesting that the phage DNA probably integrated into the bacterial genome and became a prophage. However, the compromised growth of the ΔbceQ mutant at the exponential stage by the phage treatment indicated that EPS likely protected PsJN from bacteriophage infection at the early stage of the bacterial growth.

Camelina and common pea were used to test the role of EPS in the endophytic colonization of PsJN. The wild type PsJN and the ΔbceQ mutant were injected into the leaves using a needleless syringe (Figures 6A,D). The bacterial populations were measured on day 0 and day 7 after inoculation. As shown in Figures 6B,E, the wild type PsJN and the ΔbceQ mutant had a similar level of bacterial population size on day 0. However, the ΔbceQ mutant’s population was significantly lower than the wild type PsJN on day 7 after inoculation in both plants, indicating that EPS production is important for a successful colonization of PsJN in the tested plants.

Further, we tested the bacterial colonization under plant drought stress by stopping watering immediately after the bacterial inoculation. As shown in Figure 6C, around 70% of the wild type’s population survived on day 7 under the drought stress of camelina plant relative to the bacterial population on day 7 under the normal plant growth condition. However, only less than 10% of the ΔbceQ mutant’s population survived under the plant drought stress relative to the normal plant growth condition. A similar result was also observed when the strains were tested in pea leaves (Figure 6F).

Overall, these results showed that EPS production was required for the endophytic colonization of PsJN in camelina and pea under not only the normal growth condition but also the drought stress of the host plants.