Tissue sections from lesions on diseased leaves were first disinfected with 70% ethanol, then cut and crushed using a mortar and pestle in sterile distilled water and serially diluted. One hundred microliters of each dilution was spread onto Reasoner’s 2A (R2A) agar plates (Difco, Detroit, MI, USA) and incubated at 25°C for 1 week. On the 10³-diluted plate, 30-40 colonies appeared. One white-to-yellowish and transparent colony, designated Ra3, was purified by transferring it onto a fresh plate and incubating it under the same conditions. Strain Ra3 was routinely cultured on R2A agar at 25°C and was maintained as a glycerol suspension (20%, w/v) at -70°C. Growth was assessed on R2A agar (Difco, Detroit, MI, USA), Luria-Bertani agar (LB; Difco), and potato dextrose agar (PDA; Difco) plates, and King’s medium B (KB) plates. The isolate was deposited in the Korean Collection for Type Cultures (KCTC 62486) (Korea Research Institute of Bioscience and Biotechnology, Jeollabuk-do, Korea).

The Gram stain reaction of the bacterium was determined using the staining (Smibert and Krieg, 1994) and non-staining (Buck, 1982) methods, and its oxidase activity was indicated by the development of a violet-to-purple color following the addition of 1% (w/v) tetramethyl-p-phenylene diamine (BioMérieux, Craponne, France). Fluorescent pigment was tested on King’s medium B (proteose peptone no. 3, 20 g; K₂HPO₄ · 3H₂O, 2.5 g; MgSO₄ · 7H₂O, 6 g; glycerol, 15 ml; agar, 15 g; and distilled water, 1 l). Enzyme activities, the assimilation of carbon sources, and other physiological characteristics were determined using the API 20NE, API 32 GN, and API 50 CH kits according to the manufacturer’s instructions (BioMérieux).

For the phylogenetic analysis, genomic DNA was extracted using the hexadecyl trimethyl ammonium bromide (Sigma- Aldrich, St. Louis, MO, USA) purification method as described previously (Ausubel et al., 1994). The 16S rRNA gene sequence of strain Ra3 was investigated using the method described by Weisburg et al. (1991) with the universal bacterial primers 27F and 1492R. The gyrB and rpoD genes were analyzed as described previously (Yamamoto and Harayama, 1995, 1998; Yamamoto et al., 1999). The primers UP-1E and APrU were used to amplify gyrB, while the primers 70F and 70R were used for rpoD. The partial 16S rRNA, gyrB, and rpoD gene sequences were compiled using the SeqMan software (DNASTAR, Madison, WI, USA).

The 16S rRNA, gyrB, and rpoD gene sequences were aligned using the Clustal X computer program (Thompson et al., 1997). Their phylogenetic neighbors were identified and pairwise 16S rRNA gene sequence similarities with closely related species were calculated using the EzBioCloud server (Yoon et al., 2017). Sequence similarities for 16S rRNA, gyrB, and rpoD were analyzed using the BLAST search program at the NCBI website (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Phylogenetic trees were reconstructed not only from the data sets of gyrB and rpoD, but also from the combined nucleotide sequences of these two genes (gyrB and rpoD), assuming that the analysis using a longer sequence would result in a better resolution and reliability. The 17 related taxa were obtained from GenBank and edited using the BioEdit program (Hall, 1999). These strains are listed in Table 1. Multiple sequence alignments were performed using Clustal X program. Gaps and the 5’ and 3’ ends of the alignments were edited manually in BioEdit (Hall, 1999). Evolutionary distance matrices were generated as described by Kimura (1980), and phylogenetic trees were constructed using the neighbor-joining (Saitou and Nei, 1987), maximum-likelihood (Felsenstein, 1981), and maximum-parsimony (Fitch, 1971) algorithms in the MEGA7 program (Kumar et al., 2016), with bootstrap values calculated based on 1,000 replications.

The pathogenicity was test as descripted previously (Li et al., 2007). Healthy R. asiaticus plants were used for the pathogenicity test. Bacterial colonies of the Ra3 strain were cultured on R2A agar for 48 h at 25°C and suspended in sterile distilled water. The cells were washed, re-suspended in 10 ml sterile distilled water, and counted in a hemocytometer under a light microscope (BX-50, Olympus, Tokyo, Japan), the number of cells per ml was calculated, then diluted into about 1×10⁸ colony-forming units/ml. The healthy R. asiaticus plants were inoculated with 200 μl bacterial suspensions on their stems using a syringe 10 mm below the surface and wounds were made on their stems. Control stems were treated in the same manner but only with sterile water. Each data point represents six replicates. The experiments were repeated three times with similar results. The treated plants were incubated in a moist chamber at a temperature of 25°C for 1 day, then cultured under normal conditions, and the symptoms were checked after 3 days.

Stem rot symptoms were observed on R. asiaticus three days after the inoculation with the bacterial isolate (Fig. 1). In the early stage of the disease, several black spots appeared on the stem of the infected plants. As the disease progressed, the infected stem cleaved and wilted. The infect stem lesion location shows water-soaking and brown rot. Rot symptoms developed along the outer stem at early stage and gradually into inner parts of stem. However, no disease symptoms were observed on the stems of the control R. asiaticus.

The isolated cells were Gram-negative, aerobic, rod-shaped, lacked flagella, and measured 0.8-1.0 μm wide and 1.5-2.3 μm long. Colonies grown on R2A agar at 25°C for 2 days were white-to-yellowish, transparent, circular, convex, smooth, and slimy. Growth occurred on R2A, LB, and PDA agar. Colonies were fluorescent on King’s medium B. Growth was observed at 4°C and 37°C, but not at 41°C. The strain was positive for oxidase and casein hydrolysis. The phenotypic characteristics that identified strain Ra3 as being similar to other P. marginalis strains are listed in Table 2. It showed positive results for the utilization of L-arabinose, D-galactose, gluconate (weakly), glucose, 2-ketogluconate, mannitol, propionate (weakly), sorbitol, sucrose, and trehalose, but negative results for the utilization of cellobiose and L-rhamnose. The biochemical characteristics are almost identical to those of a previously described strain (Kim et al., 2002), as shown in Table 2.

Comparative 16S rRNA gene sequence (1,400 bp) analyses showed that strain Ra3 is phylogenetically affiliated to the members of the genus Pseudomonas calculated using the EzBioCloud (Yoon et al., 2017), with sequence similarities as follows: P. grimontii CFML 97-514^T (99.85%), P. rhodesiae CIP 104664^T (99.71%), and P. marginalis ATCC 10844^T (99.70%). Since it is difficult to identify the closest species in the genus Pseudomonas using the 16S rRNA sequence alone, other gene sequences have been used as markers, such as gyrB, rpoD (Yamamoto et al., 2000), oprI (De Vos et al., 1998), and rpoB (Ait Tayeb et al., 2005). In the present study, to identify the isolated bacterial strain at the species level, the nucleotide sequences of the gyrase B (gyrB) and RNA polymerase D (rpoD) genes (911 bp and 810 bp, respectively) were obtained and compared with the GenBank database. A phylogenetic tree was obtained using the neighbor-joining method with the combined gyrB and rpoD nucleotide sequences of 13 species of Pseudomonas (a total of 17 strains). This analysis was performed under the assumption that using the gyrB and rpoD sequences in combination would give a more accurate estimate of the phylogeny than using either sequence alone. The phylogenetic position of the new isolate revealed that strain Ra3 appeared within the cluster comprising members of the genus Pseudomonas, and was grouped with several P. marginalis strains with a high bootstrap value of 100%, indicating that strain Ra3 is fully consistent with the species P. marginalis (Fig. 2). The 16S rRNA, gyrB, and rpoD gene sequences of strain Ra3 have been deposited under the accession numbers LC375532, LC379203, and LC379204, respectively.