Methylobacterium, a major component of the culturable bacterial endophyte community of wild Brassica seed

Brassica germplasm

A total of 64 accessions (49 wild and 15 landraces) of Brassica (Table S1), encompassing a diverse number of species, with a worldwide distribution, were obtained from three international genebanks, namely The United States Department of Agriculture (USDA) via The Germplasm Resources Information Network (GRIN), The Leibniz-Institute of Plant Genetics and Crop Plant Research (IPK Gatersleben) and The Nordic Genetic Resource Centre (NordGen). These accessions were imported into the Margot Forde Germplasm Centre (MFGC), New Zealand’s national genebank of grassland plants (on seed import permit no. 2015058982). A further 19 populations of wild Brassica were collected locally, within the Manawatu-Wanganui region situated in the lower half of the North Island of New Zealand. All 83 accessions were cataloged and stored at 0°C and 30% relative humidity within the MFGC.

Screening for seed-originating bacterial endophytes

Two surface disinfection protocols were developed to remove non-target microorganisms (such as saprophytic microorganisms) associated with the seed surface of the aforementioned Brassica accessions. Initially all 83 accessions were surface disinfected using the following protocol: seeds were washed for five min in 5% aqueous Tween-20^® solution (Sigma-Aldrich Inc., Auckland, New Zealand), two min in 70% ethanol, 10 min in 0.5% sodium hypochlorite, one min in 70% ethanol and rinsed three times in sterile water. To assess the efficacy of the surface disinfection protocols, 3 × 20 μL drops of water from the last rinse were plated onto nutrient agar (NA), (CM003, Oxoid Ltd., UK). Petri plates containing the NA were incubated for 2 weeks at 22 °C and inspected daily for microbial growth with the aid of a dissecting microscope (Carl Zeiss (N.Z.) Ltd., New Zealand). For those accessions where saprophytic microorganisms were initially observed, a second surface disinfection protocol was applied whereby the same procedure listed above was repeated except for one modification; seeds were immersed for 10 min in 2% sodium hypochlorite rather than a 0.5% solution. Seed were then dried on filter paper (110 mm, Thermo Fisher Scientific Ltd., Auckland, New Zealand) within a sterile environment and stored at 4 °C in the dark. To induce germination, seeds were dipped into sterile 0.2% KNO₃ solution and immediately plated onto Petri plates containing 1.5% water agar (WA), with 10 seeds per plate. Petri plates were incubated at 4 °C in the dark for 72 h to break seed dormancy. Dormant accessions (as tested for 15 accessions) resulted in either zero or poor germination without this process and were subsequently transferred to a custom-built growth chamber at 22–25 °C and a 16/8 h (light/dark) photoperiod.

Seed and the subsequent seedlings were examined daily under a dissecting microscope and those exhibiting any obvious epiphytic microbial growth were discarded. After 2–3 days, 10 clean seedlings, from each accession, were transferred to sterile tissue culture pots, 98 mm diameter (2105646, Alto Ltd., New Zealand) containing Murashige & Skoog (MS) basal salts (Murashige & Skoog, 1962) with minimal organics (Sigma-Aldrich, New Zealand), plus 3% sucrose and 1.5% agar (Ali et al., 2007). Pots were placed in the growth chamber (with the same settings as described earlier) and visually assessed every day for a month using a dissecting microscope. Plants were discarded if they showed any disease symptoms or any saprophytic microbial growth. Four clean seedlings were finally selected from each accession and subsequently dissected into two components: shoot and root. These organs were further dissected into 2–3 mm² pieces using sterile forceps and a scalpel. Ten pieces per organ type from each seedling were transferred to Petri plates containing NA. Petri plates were incubated for 3 weeks at 22 °C in the dark and checked daily under a dissecting microscope for microbial growth. Bacterial colonies arising from dissected tissue pieces were selected, sub-cultured and checked for purity. Representative bacterial isolates were then sub-cultured onto fresh NA using a sterile loop and stored in 25% glycerol at −80 °C.

Identification of seed-originating bacterial endophytes

Bacterial isolates were identified based on the PCR amplification of partial 16S rDNA gene sequences (Weisburg et al., 1991). PCR was performed directly on suspensions of each purified bacterial colony as follows: each colony was suspended in 10 μL Milli-Q^® water in a standard 0.2 mL PCR tube (Axygen^™, San Francisco, CA, USA) and frozen at −20 °C before being thawed and heated at 65 °C for 30 min. one μL of suspension was added to the PCR reaction containing five μL 10× PCR buffer, 1.5 μL MgCl₂ (50 mM), forward primer, 27F (5′-AGAGTTTGATCCTGGCTCAG,one μL, 10 μM), reverse primer R1497, (5′-CCTATATCGCCGGTAATT, one μL,10 μM), 0.4 μL dNTPS (25 mM), 0.25 μL Taq-polymerase and 39.85 μL sterile Milli-Q water to make a 50 μL PCR reaction. PCR was performed in a thermocycler (Bio-Rad C1000 Touch^TM, Bio-Rad Laboratories Inc., Hercules, CA, USA) with the following conditions: an initial step of 95 °C for 5 min was followed by 36 cycles of 94 °C for 30 s, 56 °C for 60 s, 72 °C for 90 s and a final step of 72 °C for 10 min; PCR amplification products were confirmed by electrophoresis in a 1.5% agarose gel and purified using the DNA clean & concentrator kit (Zymo Research Corporation, Irvine, CA, USA) prior to Sanger sequencing (Sanger & Coulson, 1975) (New Zealand Genomics Ltd., New Zealand). The raw sequence ab1 files were imported into the software package Geneious Prime^® version 2019.1.1 (Biomatters Ltd., Auckland, New Zealand) and were quality trimmed using an error probability of 0.05. Those sequences with a region of high quality greater or equal than 600 bp were kept and annotated using the BLASTn program against the NCBI rRNA database. The sequences were aligned using the multiple alignment program MAFFT (Katoh & Standley, 2013), and a Maximum Likelihood phylogenetic tree was generated using the software Mega X (Kumar et al., 2018) using a general time reversible model and validated by 100 bootstrap cycles. All sequence data were deposited in GenBank under file SUB6483552: MN629046–MN629135. Simpson`s diversity index (Simpson, 1949) was used as a quantitative measure to reflect how many different bacterial species there were in the seed-originating microbial community of the Brassica accessions sampled.

Assessing plant growth promotion

Two isolates of Methylobacterium, namely Methylobacterium fujisawaense (isolate B82) and Methylobacterium phyllosphaerae (isolate B64), were selected due to their high tissue colonization rate in their original host plants. The bacteria were plated on NA and incubated for 2 weeks at 22 °C. For each isolate, cells were scraped from the Petri plates using a loop and transferred to an aqueous Tween-20^® solution. Concentrations were adjusted to 10⁹ cells per mL using a hemocytometer. Oilseed rape, cv. King was selected as the novel host plant. Seeds were surface disinfected, as described earlier, and placed on a filter paper under a laminar flow cabinet to dry. They were then transferred to Petri plates containing 2% WA and incubated at 22 °C in a custom-made lighting room with 18/6 h (light/dark photoperiod) to initiate germination. The root tip of each seedling was excised with a sterile scalpel, dipped into the prepared bacterial suspension and transplanted into sterile plastic pots (7 × 15 cm) containing autoclaved potting mix (50% fine bark, 12.5% compost and 25% pumice plus nutrient, gypsum and Agri-lime). Control seedlings, after excising the root, were dipped in sterile water containing one drop of Tween-20^® per liter. Pots were watered equally, and lids placed on top. Pots were transferred to a plant growth chamber (A1000, Conviron Asia Pacific Pty Ltd., Melbourne, Australia) set at 18 °C with a 16/8 h (light/dark) photoperiod. The experiment was laid out in a completely randomized design with eight replications. Each experimental unit comprised five pots/plants. After one month, plants were removed, and all soil debris cleaned by washing under a water tap. Each plant was placed on a filter paper for 8 h at room temperature (20–25 °C) to completely dry. The seedlings were weighed and the mean weight of five plants in each experimental unit were used for analysis of variance (ANOVA) using SPSS software (IBM^© SPSS^© Statistics, version 24).

Dual culture test

The antifungal activity of a representative isolate of each bacterial species was assessed in vitro (dual culture) against the target pathogen, L. maculans. The L. maculans strain Lm145 utilized was a virulent pathogen of Brassica sp., previously isolated from a swede crop in New Zealand (Lob, 2014). Bacterial endophytes, previously stored at −80 °C, were thawed, streaked onto NA and incubated for 2 weeks at 22 °C. For each bacterial isolate a cell suspension (10⁹ cells per mL) was prepared and 50 μL streaked in a straight line across the center of a Petri plate. Control treatments were streaked with only sterile water mixed with one drop of Tween-20^® per liter. Petri plates were incubated for 2 weeks at 22 °C and then two mycelial plugs (5 mm diameter) were taken from the actively growing region of a 2-week old L. maculans culture and placed 25 mm from the edge of each Petri plate opposite each other; there were 10 replicate plates for each treatment. When the two fungal colonies placed on opposite sides had grown sufficiently to close the gap between them, the distance between the bacterial colony and the L. maculans colony was measured using a digital caliper (Mitutoyo Corporation, Kawaski, Kanagawa, Japan). The inhibition zone was rated using a 5 point scaling system: 4 (very high), inhibition zone >10 mm; 3 (high), imbibition zone 5–10 mm; 2 (medium), inhibition zone <5 mm; 1 (low), L. maculans growth stopped at the bacterial streak line; 0 (zero), no inhibition zone with L. maculans typically growing over the bacterial streak (Hammoudi et al., 2012).

Seed-originating bacterial isolates

In total, 54 accessions (44 wild and 10 landrace), out of the total 83 accessions surface disinfected and sown, resulted in symptomless plants when germinated and subsequently grown on MS medium. The remaining 29 accessions all exhibited epiphytic fungal growth, with most colonies identified as Alternaria sp. These plants were all destroyed by autoclaving. After incubation, bacterial colonies were detected from 48 symptomless wild Brassica accessions (38 wild and 10 landrace). Six accessions did not result in any bacteria being isolated.

Identification of seed-originating bacteria

A sample of 90 bacterial isolates, representing the morphological diversity of all the observed bacterial colonies that developed from dissected Brassica tissue, were sequenced to determine their species identity. 16S rDNA gene sequencing identified 19 different bacterial species belonging to three phyla, namely Actinobacteria, Firmicutes and Proteobacteria (Fig. 1; Table 1). According to the phylogenetic tree (Fig. 1) the most frequently isolated species were Methylobacterium fujisawaense (50 isolates), Stenotrophomonas rhizophila (10 isolates) and Pseudomonas lactis (9 isolates). Three species were identified from the Actinobacteria, namely Kocuria palustris, Micrococcus aloeverae, and Plantibacter flavus while two species were identified from the Firmicutes phylum, namely Bacillus mycoides and Paenibacillus hordei.

Simpson’s diversity index recorded a value of 0.74 indicating that these wild and landrace Brassica accessions contained a high diversity of seed-originating bacteria. Nevertheless, Methylobacterium spp., predominantly Methylobacterium fujisawaense and the closely related species Methylobacterium phyllosphaerae, Methylobacterium oryzae and Methylorubrum extorquens constituted 56% of the bacteria isolated and were common in 77% of the Brassica accessions screened (38 accessions: 28 wild and 10 landrace). These were also distributed among multiple Brassica species including Brassica barrelieri, Brassica elongate, Brassica gravinae, Brassica indica, Brassica juncea, Brassica napus, Brassica nigra and Brassica rapa sourced from five continents, namely Africa, Asia, Australasia, Europe and North America.

Plant growth promotion

Oilseed rape plants that were inoculated with two isolates of Methylobacterium showed a significant (P < 0.01) increase in their fresh weight when compared to the uninoculated control plants (Fig. 2). Plants inoculated with Methylobacterium fujisawaense (B82) had a mean weight of 1.33 g/plant, while those inoculated with Methylobacterium phyllosphaerae (B64) displayed a mean weight of 0.88 g/plant compared to a mean weight of 0.69 g/plant for the uninoculated control plants (Fig. 2).

Dual culture test

The results from the dual culture assays indicate that some of the bacterial isolates showed antagonistic behavior towards L. maculans. The strongest inhibition effect was observed with isolates of Methylobacterium fujisawaense and Methylobacterium phyllosphaerae, N. resinovorum, Pseudomonas lactis, Plantibacter flavus and Stenotrophomonas rhizophila that all showed a high, clear inhibition zone between the edge of the bacterial streak and the pathogen (Fig. S1). However, L. maculans was not affected by K. palustris, Sphingomonas yantingensis, Sphingomonas insulae, Bacillus mycoides or Brevundimonas vesicularis; with no inhibition zone produced.