Various gymnosperms and angiosperms, including Metasequoia glyptostroboides, Ginkgo biloba, Taxus brevifolia, Pinus densiflora, Salix babylonica, and Salix chaenomeloides were sampled at Yeungnam University, Gyeongsan, Republic of Korea in October 2014. Leaves, fruit, seeds, and cones (Table 1) were collected, placed in plastic bags, and brought to the laboratory for EB isolation.

EB were isolated according to standard procedures (De Oliveira Costa et al., 2012) with minor modifications. Plant parts were washed under running tap water. Visibly injured organs were discarded. The samples were dried and weighed on a digital balance. Tissues were sterilized by exposure to 70% v/v alcohol for 60 s, 2% sodium hypochlorite for 180 s, and 100% ethanol for 30 s. The tissues were then rinsed several times with sterile distilled water (SDW) and dried with sterile filter papers. The plant parts were then ground with sterile 3 ml/g of aqueous NaCl solution (0.9% w/v) using a sterile mortar and pestle. The tissue extracts were incubated at 28°C for 3 h to release all of the EB.

The tissue extracts were diluted to 10⁻¹ and 10⁻² with sterile NaCl solution. They were then applied with a stainless steel spreader onto Petri dishes containing 25% yeast extract, nutrient broth and agar (YNA), 25% nutrient agar, de Man, Rogosa and Sharpe agar, or Luria-Bertani agar media to obtain single bacterial colonies. The plates were incubated for up to 15 days at 28°C. All colonies were counted and reported as the number of colony-forming units per gram fresh tissue [cfu/g fresh weight (FW)]. Single colony isolation was repeated at least three times to obtain pure EB isolates. Colony form, margin, color, and elevation were recorded. The selected EB were cultured in various media broths and stored in 50% glycerol at −80°C until use.

The isolated EB were identified with 16S rRNA gene sequencing. The sequencing data, alignment, and phylogenetic trees were analyzed by Molecular Evolutionary Genetics Analysis (MEGA) 6 software version 6.0. The phylogenetic trees were also analyzed by the neighbor-joining method (Susilowati et al., 2015; Tamura et al., 2013).

Nine Xcc strains, including three wild-type and six streptomycin-resistant mutants (Hyun et al., 2012) were used in this study. The wild-type (XccW1 [Xcc 19-18], XccW2 [Xcc 10-5], and XccW3 [Xcc 53-4]) and streptomycin-resistant mutant (XccM4 [Xcc 27-9], XccM5 [Xcc 8-4], XccM6 [Xcc 57-2], XccM7 [Xcc 57-5], XccM8 [Xcc 27-13], and XccM9 [Xcc 27-12]) strains were kindly provided by Dr. Hyun of the Citrus Research Station, National Institute of Horticultural and Herbal Science, Rural Development Administration, Jeju, Korea. The wild-type Xcc were grown in yeast extract-nutrient broth (YNB) medium. The streptomycin-resistant Xcc strains were incubated in YNB supplemented with 50 mg/ml streptomycin. All Xcc were stored in 50% glycerol at −80°C until use.

The EB isolates were screened for antagonistic activity against the nine strains of Xcc following standard procedures (Roh et al., 2009) with minor modifications. Screening was performed by culturing 134 EB isolates in 96-well plates with 10 μl individual bacterial strain grown in YNB and the plates were incubated for 36 h at 28°C. Then the cultured bacteria were again grown on square plates (size as 127.94 × 85.50 × 16.25 mm, SPL Life Sciences Co., Ltd., Pocheon, Korea) using 25% YNA (1.5% agar) media following incubation for 36 h at 28°C. The EB were then killed by placing 1 ml chloroform on the square plate lids and inverting the plates for 15 min. The lids were then removed and the plates were left to stand for 20 min to vent any chloroform residue, then the plates were treated with UV light for 15 min to kill the bacteria completely. Simultaneously, 30 μl aliquots of Xcc cultures that had been grown overnight in YNB media at 28°C were suspended in 10-ml fresh YNB media (0.75% agar) at 55°C, mixed gently, then poured slowly and uniformly on the square plates. After the agar solidified, all the plates were sealed and incubated at 28°C for 24 h. Inhibition zone diameters were measured using an electronic digital caliper (M500-182M, Hando, Cheonan, Korea) with a 1/100 mm precision.

Further analysis for the antagonistic activity was done using TbL-22 (B. thuringiensis) and TbL-26 (B. thuringiensis) as the isolates with the highest antagonistic activity along with TbS-8 (B. altitudinis) and TbL-19 (B. altitudinis) as the isolates of another Bacillus species against all of Xcc strains. The above-described method was used the same way using all Xcc strains and using Petri dish (90 mm diameter) instead of square plates.

After being sub-cultured on YNA plates for 24 h, single colonies of the antagonistic EB were added to 200 ml YNB in 1-liter Erlenmeyer flasks with continuous shaking at 180 rpm at 28°C for 48 h. The culture broths were centrifuged at 3,500 rpm for 15 min at 4°C. The supernatant was collected, mixed with equal volumes of ethyl acetate and incubated overnight. Ethyl acetate was selected for the metabolite extraction process for its low boiling point and moderate polarity. Simultaneously, the bacterial pellets were suspended in 20 ml methanol, placed in a sonicating bath (WUC-A03H, Daihan Scientific Co., Wonju, Korea) for 15 min to break the cells and incubated overnight. The ethyl acetate layer was then collected in a separating funnel. The methanol extract was separated from the bacterial debris by centrifugation at 3,500 rpm for 15 min at 4°C. Both extracts were dried at 50°C in a rotary evaporator (A-1000S, EYELA, Tokyo, Japan). The dried extracts were then removed from the evaporation flask, dissolved in 2 ml methanol, and air-dried in tubular glass vials. The ethyl acetate extract was measured and dissolved in 5% dimethyl sulfoxide (DMSO) in methanol to make 0.1 g/ml. Total 50-μl aliquots of this dilution were used to the well diffusion assay. The antagonistic activities of the EB against the Xcc strains were determined by measuring the inhibition zone diameters.

The MIC and MBC of the ethyl acetate extract of TbL-22 against Xcc were determined by the two-fold dilution method (Patra et al., 2015) with minor modifications. Five percent DMSO was used both as a control and a solvent to homogenize the ethyl acetate extract for the MIC and MBC tests. Various dilutions of the ethyl acetate extract of TbL-22 in YNB media were prepared. The lowest concentration of TbL-22 at which there was zero visible Xcc growth was taken to be the MIC and was recorded in mg/ml. The MIC and the next higher concentration were selected, spread on YNA plates and incubated at 28°C for 24 h. The lowest concentration of TbL-22 at which no pathogenic bacteria grew on the 25% YNA plates was considered the MBC and was expressed in mg/ml.

The effects of the ethyl acetate extract of TbL-22 on Xcc cell morphology were determined by SEM analysis. Both the 5% DMSO control and the samples treated with the ethyl acetate extract of TbL-22 were applied to the four Xcc strains and observed under SEM using a slightly modified version of the procedure described by Bajpai et al. (2009). The four different Xcc strains used for SEM analysis were XccW1, XccM4, XccM6, and XccM8, which were selected based on their genotypic classification. Overnight bacterial culture samples were washed gently with 0.05 M phosphate buffer solution (pH 7.4) and fixed with 2.5% v/v glutaraldehyde in SDW. The specimens were dehydrated with sequential exposures to ethanol ranging from 50 to 100%. The ethanol was then replaced with tert-butyl alcohol. The specimens were sputter-coated with platinum in an ion coater for 120 s then examined under an SEM (S-4100, Hitachi, Tokyo, Japan).

Pathogenic wild-type XccW2 were grown at 28°C in YNB media. Overnight bacterial cell cultures were harvested at 4°C, centrifuged at 3,500 rpm for 15 min, and suspended in SDW to make a concentration of 0.4 OD_600nm determined by spectrophotometer (ASP-3700, ACTGene, Piscataway, NJ, USA). For the detached leaf assay (Brown and Soepena, 1994), fully expanded immature (4-week-old) leaves were excised from Hwanggeum hyang, a highly canker-susceptible cultivar generated from the cross between Citrus unshiu × C. sinensis and C. unshius. The leaves were infiltrated with a 0.1 ml bacterial suspension using a 1-ml needleless syringe. The bacterial suspensions were prepared by mixing 0.4 OD_600nm XccW2 either with 0.4 OD_600nm EB suspension or with SDW. The infiltrated leaves were placed on three layers of sterile tissue paper moistened with SDW in a humid box and incubated in a growth chamber at 28°C. The infiltrated leaves were photographed, and any canker lesions that developed were measured 7 days post-infiltration (dpi) (Martínez-Medina et al., 2013).

To determine whether TbL-22 and TbL-26 could cause disease in citrus plants, bacterial suspensions at 0.4 OD_600nm concentration were infiltrated into Hwanggeum hyang leaves as described by Brown and Soepena (1994). The pathogenic bacterium XccW2 and SDW were used as a positive and negative control, respectively. Citrus canker disease symptoms were assessed for up to 7 dpi.

All data were expressed as the mean ± standard deviation by measuring four independent replicates per experiment. Statistical analysis of the results was conducted by one-way ANOVA followed by the Duncan’s multiple range test at P < 0.05 using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

A total of 134 morphologically distinct EB were isolated from M. glyptostrobodies, G. biloba, T. brevifolia, P. densiflora, S. babylonica, and S. chaenomeloides. They were obtained from the leaves (73 isolates, 54.48%), fruit (49 isolates, 36.57%), seeds (8 isolates, 5.97%), and cones (4 isolates, 2.98%) (Table 1). The tissue-specific cell counts (CFU) per gram FW were in the range of 1.8 × 10²-6.0 × 10⁴ cfu/g in the fruits, 1.8 × 10³-1.5 × 10⁵ cfu/g in the leaves, 7.2 × 10³ cfu/g in the seeds, and 1.2 × 10³ cfu/g in the cones (Table 1).

In the primary screening, all EB isolates were tested for antagonistic activity against XccW1. Eighteen of the EB efficacious against XccW1 were identified by 16S rRNA gene sequencing analysis (Table 2). Among them, TbL-22 and TbL-26 exerted the highest antagonistic activity against all Xcc, so they were selected for further study (Table 3). Phylogenetic trees were constructed using the 16S rRNA sequences of TbL-22 and TbL-26 and those of their closest relatives. The phylogenetic tree analysis revealed that TbL-22 and TbL-26 had 99% and 100% identity with the 16S rRNA gene sequences of B. thuringiensis Bt 407 and B. thuringiensis IAM12077, respectively (Fig. 1A and B).

In Xcc cultures, TbL-22 and TbL-26 exhibited the highest antagonistic activity with inhibition zones of 20.64-15.33 and 17.34-11.65 mm, respectively (Table 3, Fig. 2). TbL-22 had greater activity than TbL-26 against all Xcc. Only the ethyl acetate extract of TbL-22 was antagonistic against all Xcc strains (Table 4, Fig. 3). The extract of TbL-22 exerted the highest antagonistic activity of all, and produced inhibition zones of 15.31 ± 2.08 mm and 19.37 ± 3.17 mm against XccW3 and XccM6, respectively. The extract of TbL-26 produced inhibition zones of 13.81 ± 1.05 mm and 18.08 ± 4.69 mm against XccW3 and XccM6, respectively, but had no antagonistic activity against XccW2 (Table 4, Fig. 3).

Since the extract of TbL-22 had the highest antagonistic activity against all Xcc, it was selected for the determination of MIC and MBC. The values of MIC and MBC for the ethyl acetate extract of TbL-22 ranged from 0.625 to 5.00 mg/ml and from 1.25 to >5.00 mg/ml, respectively (Table 5). The lowest MIC and MBC values were found for XccW2 (0.625 mg/ml and 1.25 mg/ml, respectively) (Table 5).

The effects of the ethyl acetate extract of TbL-22 on cell morphology were observed for XccW1, XccM4, XccM6, and XccM8 (Fig. 4). XccW1 was selected to represent the wild-type, and XccM4, XccM6, and XccM8 were selected as representatives of each mutant on the basis of their genotypic classification. All Xcc treated with the ethyl acetate extract presented with ruptured cell surfaces. They were elongated, shrunken, lysed, and swollen. In contrast, the control Xcc cell surfaces were regular, uniform, and smooth (Fig. 4).

TbL-22 and TbL-26 were tested for in vivo biocontrol efficacy against Xcc. In the detached leaf assay, all citrus leaves infiltrated with XccW2, TbL-22, and TbL-26 showed disease symptoms. Nevertheless, the infected areas of the leaves infiltrated with TbL-22 and TbL-26 were smaller than those of the leaves treated with XccW2 alone by 64.05% and 47.76%, respectively (Fig. 5A and B). After 7 dpi, XccW2 infiltration alone caused the widest necrotic lesion (10.68 ± 1.34 mm), followed by XccW2 with TbL-26 (5.58 ± 0.51 mm) and XccW2 with TbL-22 (3.84 ± 0.27 mm) (Fig. 5C).

The lesions that developed on the leaves infiltrated with XccW2 were usually visible on both surfaces and appeared in the form of water-soaked margins surrounded by yellow rings or haloes (Fig. 6A). Leaves infiltrated with TbL-22 and TbL-26 alone, however, did not display any visible toxic effects or canker symptoms.