Development of a New Semi-Selective Lysine-Ornithine-Mannitol-Arginine-Charcoal Medium for the Isolation of Pantoea Species from Environmental Sources in Japan

Although Pantoea species are widely distributed among plants, water, soils, humans, and animals, due to a lack of efficient isolation methods, the clonality of Pantoea species is poorly characterized. Therefore, we developed a new semi-selective medium designated ‘lysine-ornithine-mannitol-arginine-charcoal’ (LOMAC) to isolate these species. In an inclusive and exclusive study examining 94 bacterial strains, all Pantoea strains exhibited yellow colonies on LOMAC medium. The performance of the medium was assessed using Pantoea-spiked soils. Percent average agreement relative to the Api20E biochemical test was 97%. A total of 24 soil spot samples and 19 plant types were subjected to practical trials. Of the 91 yellow colonies selected on LOMAC medium, 81 were correctly identified as Pantoea species using the biochemical test. The sequencing of 16S rRNA (rrs) and gyrB from these isolates confirmed that Pantoea agglomerans, P. vagans, P. ananatis, and P. deleyi were present in Japanese fields. A phylogenetic analysis using rrs enabled only the limited separation of strains within each Pantoea spp., whereas an analysis using gyrB revealed higher variability and enabled the finer resolution of distinct branches. P. agglomerans isolates were divided into 3 groups, 2 of which were new clades, with the other comprising a large group including biocontrol strains. P. vagans was also in one of the new clades. The present results indicate that LOMAC medium is useful for screening Pantoea species. The use of LOMAC medium will provide new opportunities for identifying the beneficial properties of Japanese Pantoea isolates.

Pantoea is a genus of Gram-negative bacteria of the family Enterobacteriaceae, recently separated from the genus Enterobacter. The genus Pantoea includes at least 20 species, such as P. agglomerans, P. vagans (formerly a P. agglomerans strain), P. ananatis, P. deleyi, and P. eucalyptii (25). Members of Pantoea are motile, non-encapsulated, non-spore forming rods with peritrichous flagella, and are typically yellow pigmented. Pantoea are abundant in plant and animal products, arthropods and other animals, water, soil, dust and air, and occasionally in humans. Pantoea species exhibit both deleterious and beneficial characteristics. For example, although Pantoea species are known to cause crop diseases and disorders in exposed individuals via the inhalation of organic dusts (5-7, 16, 18), they also produce substances effective in the treatment of various cancers in humans and animals, suppress the development of various plant pathogens via antibiotic production and/or competition, and exhibit bio-fertilizer and bio-remediation properties (6).
Previous studies reported the isolation of unique Pantoea strains from environmental sources. Son et al. in Korea, Malboobi et al. in Iran, and Sulbaran et al. in Venezuela demonstrated that P. agglomerans strains isolated from soil exert beneficial effects on crops (6,14,15,23,24). In greenhouse and field trials, Malboobi et al. showed that P. agglomerans promoted the growth of potato plants (14,15). Kageyama et al. isolated new Pantoea spp. from fruits and soils in Japan and demonstrated that these Japanese species were phylogenetically distant from other Pantoea species (3,11). Japanese researchers also reported the efficacy of LPS from P. agglomerans for the treatment of human cancers. Kasugai et al. intradermally administered LPS in combination with transarterial intermittent chemotherapy to treat patients with advanced gastric cancer with multiple liver metastases (12).
Despite the apparent medical and agricultural significance of Pantoea spp., limited information is currently available on their distribution and prevalence, which is partly due to the lack of methods for efficient isolation and enumeration in the presence of competing organisms. Only a few media for detecting Pantoea spp. have been reported to date. Lysine-Ornithine-Mannitol (LOM) agar was developed in 1981 for isolating Enterobacter agglomerans (4); however, this medium was developed for testing human stool samples. PA 20 semi-selective medium was developed in 2006 for the isolation and enumeration of P. ananatis from plant material (10). Non-selective agar media, such as Nutrient agar and LB, are occasionally used to isolate Pantoea spp. The ability to produce a yellow pigment is used to identify Pantoea spp. on nonselective agar media, such as LB and Trypticase soy agar (5,8,9). However, the detection and isolation of Pantoea spp. from environmental sources (1,9) will require a new semiselective agar medium. Therefore, the purpose of the present study was to develop a semi-selective medium for the isolation and enumeration of Pantoea spp. in the presence of competing organisms frequently found in soils and plants.

Semi-selective agar medium
The ingredients of lysine-ornithine-mannitol-arginine-charcoal (LOMAC) agar medium were placed into two groups. Solution A was prepared by adding the following (L -1 ) to water: 3 g of yeast extract (Difco Laboratories, Detroit, MI, USA), 2 g of sodium chloride (Wako Pure Chemical Industries, Osaka, Japan), 2 g of magnesium sulfate (Wako Pure Chemical Industries), 0.05 g of sodium pyruvate (Wako Pure Chemical Industries), 1 g of soy peptone (Conda, Madrid, Spain), 5 g of L-lysine hydrochloride (Peptide Institute, Osaka, Japan), 6.5 g of L(+)-ornithine hydrochloride (Peptide Institute), 5 g of L-arginine hydrochloride (Peptide Institute), 0.3 g of bromothymol blue (Sigma Aldrich, St. Louis, MO, USA), 13.5 g of agar (SSK Sales, Shizuoka, Japan), and 2 g of charcoal (Serachem, Hiroshima, Japan). pH was adjusted to 6.5. Solution A was autoclaved at 121°C for 15 min and cooled to 50°C. Solution B was prepared by adding the following (50 mL -1 ) to water: 5.25 g of mannitol (Wako Pure Chemical Industries), 0.03 g of vancomycin hydrochloride (Wako Pure Chemical Industries), and 0.016 g of amphotericin B (Wako Pure Chemical Industries). Fifty milliliters of solution B was filter-sterilized and then mixed with 1 L of solution A. Charcoal was used as an absorbent against toxic chemicals to bacteria. Vancomycin and amphotericin B were used as inhibitors of Gram-positive bacteria and fungi, respectively. Media were designed for mannitol-positive, lysine-negative, ornithine-negative, and argininenegative species, including Pantoea spp., to yield intensely yellow colonies (8). Mannitol-, lysine-, ornithine-, and arginine-negative species yield colorless colonies. Species that are mannitol-positive and either lysine-, ornithine-, or arginine-positive yield greenish-blue colonies. Species that are mannitol-negative and either lysine-, ornithine-, or arginine-positive yield colorless colonies that turn greenish-blue after more than 24 h.

Plating efficiency tests
A total of 86 bacterial and 8 fungal strains were examined (Table  1). These included 69 Gram-negative rods, including 13 Pantoea spp., 14 Gram-positive cocci, and 3 Gram-positive rods. These strains were sub-cultured on non-selective medium (Tryptic Soy Agar [TSA]; Difco Laboratories) at 35±2°C for 24 h. An overnight TSA culture of each bacterial colony was streaked onto LOMAC and LOM plates (4), which were incubated at 35±2°C and examined after 24 h for the presence or absence of yellow colonies.
The recovery of Pantoea species using LOMAC medium was evaluated based on the efficiency of colony formation. Bacterial suspensions from fresh colonies grown on TSA were adjusted in sterile saline to an optical density at 660 nm (OD 660 ) of 0.5 (ca. 1.5×10 8 CFU mL -1 ). Bacterial suspensions were serially diluted 10-fold, and 0.1-mL aliquots of each dilution were plated on the tested plate media. The recovery percentage was calculated from the ratio of the mean colony counts on the test medium and on nonselective TSA as a reference.
Verification of LOMAC agar medium for the cultivation of Pantoea spp. from environmental soils.
The effectiveness of the LOMAC plate for bacterial recovery was demonstrated in soils (1 g) spiked with 1:10 5 dilutions of P. agglomerans NBRC 102470 (1.5×10 8 CFU mL -1 ). Soils spiked with P. agglomerans were added to sterile saline (20 mL) and vortexed for 30 s in a mixer. The supernatant was serially diluted 1:10 in sterile saline, and 0.1 mL of the suspension was spread on LOMAC agar. The plates were incubated for 24 to 48 h at 35±2°C. After the incubation, all colonies obtained were streaked onto TSA medium and incubated at 35±2°C for 24 h. Isolates were identified by biochemical testing using the Api20E system (Sysmex bioMérieux, Tokyo, Japan).

Sample collection and practical trials
Samples of plants and environmental soils were obtained from 26 areas in Japan. Samples of soil (1 g) were added to sterile saline (20 mL) and vortexed for 30 s in a mixer. Similarly, plant samples (1 g) were added to sterile saline (10 mL) and homogenized using a mortar and pestle. After filtering using a cell strainer (40 μm), the flow through fraction was serially diluted 1:10 in sterile saline, and 0.1 mL of the suspension was spread on LOMAC agar medium and non-selective TSA medium. After an incubation at 35±2°C for 24 to 48 h, bacterial colonies with a yellow color were selected, passaged onto TSA medium, and incubated at 35±2°C for 24 h. After a second incubation, the isolates were identified by biochemical testing using the Api20E system.

PCR and phylogenetic analysis of sequencing data
Bacterial DNA was extracted using a QIAamp UCP Pathogen Mini kit (Qiagen KK, Tokyo, Japan). PCR amplification of the housekeeping genes 16S rRNA (rrs) and gyrB was performed using the following primer sets: 16S-8F and 16S-1492R for rrs, and gyr-320 and rgyr-1260 for gyrB (20). PCR targeting for rrs encoding 16S rRNA was performed with the ExTaq (Takara Bio, Shiga, Japan) enzyme under the same conditions as those previously described (20), except for an annealing temperature set to 49°C. PCR targeting for gyrB encoding partial GyrB was performed with initial denaturation and activation of the ExTaq enzyme at 95°C for 5 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 50°C for 45 s, elongation at 72°C for 60 s, and final elongation at 72°C for 7 min. Positive PCR amplification was verified electrophoretically using 5 μL of each reaction loaded onto a 1.5% agarose gel. PCR products were verified by DNA sequencing. Briefly, the PCR amplicon was purified with the NucleoSpin Gel and PCR Clean-up kit (MACHEREY-NAGEL GmbH & Co. KG, Duren, Germany) and subjected to DNA sequencing. In 16S rRNA sequencing, additional primers, 16S-609R and 16S-533R, were used to achieve the complete coverage of the amplicon (20). In GyrB sequencing, DNA sequences were elucidated by the dideoxy termination method employing the same primers used for PCR amplification. Nucleotide sequences were searched for homology by BLAST screening against the GenBank databases. DNA sequences were aligned using ClustalW, and phylogenetic trees were generated based on partial gyrB sequences. Sites exhibiting alignment gaps were excluded from the analysis. NJplot program (19), version 2.3, was used to calculate evolutionary distances and infer trees based on the minimum evolution (ME) method using the maximum composite likelihood formula. The nodal robustness of the inferred trees was assessed using 1,000 bootstrap replicates.
Nucleotide sequence accession number.
The nucleotide sequences of the isolates in Japan reported here have been deposited in the EMBL/GenBank/DDBJ databases and assigned the following accession numbers: LC422596 and LC422697 to LC422728 for the nucleotide sequence of gyrB; LC438406 to LC438434 for the nucleotide sequence of rrs.

Growth and selectivity tests
To assess the selectivity of LOMAC medium for Pantoea strains and the growth of these strains on the medium, 86 bacterial strains, including P. agglomerans NBRC 102470, 6 isolates of P. agglomerans, 2 isolates of P. ananatis, P. brenneri ES153, P. deleyi ES168, 2 isolates of P. vagans, and 8 fungal strains were streaked on agar plates (Table 1). All Pantoea strains tested formed yellow colonies on LOMAC medium after being incubated for 24 h. However, 5 Gramnegative rods also formed yellow colonies on LOMAC medium. Acinetobacter lwoffii 85, Candida kefyr 116, Cryptococcus neoformans 105, and 14 Gram-positive cocci and 3 Grampositive rods did not grow on the medium. The remaining isolates, including 50 Gram-negative rods and 6 fungi, formed colonies of various colors, such as blue, green, blue-white, white, whitish-yellow, clear, brownish-yellow, and red, but Red Yellow did not form yellow colonies. These results indicated that LOMAC medium is not only semi-selective for Pantoea strains, but also enables the differentiation of strains based on colony color. Similarly, the selectivity of LOM medium for Pantoea strains and growth of the strains on the medium were tested ( Table 1). All Pantoea strains tested formed yellow colonies on LOM medium after being incubated for 24 h. However, 30 Gram-negative rods and 2 fungi also formed yellow colonies on LOM medium. The remaining isolates, including 25 Gram-negative rods and 4 fungi, formed colonies of various colors, but did not form yellow colonies. There were no strains that were present on LOMAC medium but not on LOM medium. These results indicated that LOM medium is equal to LOMAC medium for selectivity to Pantoea strains, but not superior to LOMAC for the differentiation of strains based on colony color.

Evaluation of LOMAC agar medium for the cultivation of Pantoea spp. in soils
Bacterial recovery from soils was assessed using 3 different soils spiked with P. agglomerans NBRC 102470 (Table 2). LOMAC agar plates enabled the growth of 99 colonies, 10 of which were colonies with a yellow color. Fifty-five colonies were obtained from soil A; five of these were colonies with a yellow color and correctly identified as Pantoea spp.3 using Api20E. The other 50 colonies with non-yellow colors were not identified as any Pantoea species (Table 2). Soil B yielded 4 colonies with a yellow color. Only 1 colony was identified as Pantoea spp.3. The other 2 colonies were identified as Citrobacter youngae and the remaining colony as Leclercia adecarboxylate. Soil B also yielded 26 colonies with non-yellow colors. None of them were identified as Pantoea species. Similarly, Soil C yielded 1 colony with a yellow color, which was identified as Pantoea spp. 3. Soil C yielded 13 colonies with non-yellow colors, none of which were identified as Pantoea species (Table 2). Average percent positive and negative predictive values were 70% (7/10) and 100% (89/89), respectively. Overall agreement was 97% (96/99), indicating that Pantoea species were successfully isolated on the medium and that the colonies with a yellow color were instantaneously distinguishable as Pantoea species.

Sample collection and practical trials for isolating Pantoea species
A total of 26 trials for isolating Pantoea species were performed using 24 spots of soil samples and 19 samples of plants obtained from geographically diverse regions of Japan, such as Nagano, Fukuoka, Chiba, and Hokkaido (Table 3). All samples were tested with LOMAC and TSA. LOMAC agar typically generated yellow colonies (Fig. 1a), whereas TSA medium did not have the ability to isolate Pantoea species producing a yellow pigment (Fig. 1b).
A total of 797 yellow colonies were generated on LOMAC agar (Table 3). Among these colonies, 91 were sub-cultured on TSA medium. Eighty-one out of the 91 colonies were identified as Pantoea spp. using the Api20E test. One isolate was identified as Pantoea spp.

PCR and sequencing results
In further analyses of the 81 Pantoea spp. colonies, 34 genomic DNAs were randomly selected and subjected to PCR amplification targeting the rrs gene. In 29 out of 34 genomic DNAs, PCR analyses yielded 1503-bp amplicons of the expected size.
DNA sequence analyses revealed that all of the sequences of the 20 Pantoea spp. 3 isolates were 100% identical not only to one another, but also some types of Pantoea strains, such as P. agglomerans ATCC27155, P. deleyi LMG24200 (20), and P. vagans LMG24199 (20); the rrs sequences of 8 isolates of Pantoea spp. 2 were categorized into 3 types of sequences for their highest identity, i.e., the sequences of the rrs of 5 isolates of Pantoea spp. 2 were 100% identical to P. ananatis ATCC 27966 [P. ananatis type]; the sequences of 2 Pantoea spp. 2 isolates showed the highest identity to that of Erwinia aphidicola ATCC 27992 (ranging from 99% to 100% nucleotide [nt] identity) (Erwinia type); the sequence of one Pantoea spp. 2 isolate was 100% identical to P.  agglomerans ATCC 27155 (P. agglomerans type); and the sequence of one Pantoea spp. 4 isolate showed the highest identity to that of Enterobacter cloacae ATCC 13047 (99% nt identity) (E. cloacae type).
In all 34 genomic DNAs, PCR targeting the gyrB gene yielded 970-bp amplicons of the expected size. DNA sequence analyses and a BLAST search of the gyrB amplicons revealed that the gyrB sequence of Pantoea spp. 1 isolate was nearly identical to that of E. toletana (85% nucleotide [nt] identity); the gyrB sequences of 9 isolates of Pantoea spp. 2 were categorized into 3 types of sequences for their highest identity, i.e., the sequence of one isolate of Pantoea spp. 2 showed the highest identity to that of P. vagans (97% nt identity) (P. vagans type), the sequences of 6 isolates of Pantoea spp. 2 showed the highest identity to that of P. ananatis (ranging between 91 and 100% nt identity) (P. ananatis type); the sequence of one isolate of Pantoea spp. 2 showed the highest identity to that of E. rhapontici (90% nt identity) (Erwinia type), and the sequence of one isolate of Pantoea spp. 2 showed the highest identity to that of E. tasmaniensis (87% nt identity) (Erwinia type); the gyrB sequences of 23 isolates of Pantoea spp. 3 were categorized into 4 types of sequences for their highest identity, i.e., the sequences of 15 isolates of Pantoea spp. 3 showed the highest identity to that of P. agglomerans (ranging between 95 and 99% nt identity) (P. agglomerans type), the sequences of 5 isolates of Pantoea spp. 3 showed the highest identity to that of P. vagans (ranging between 97 and 100% nt identity) (P. vagans type), the sequences of 2 isolates of Pantoea spp. 3 showed the highest identity to that of P. deleyi (99% nt identity) (P. deleyi type), and the sequence of one isolate of Pantoea spp. 3 showed the highest identity to that of P. brenneri (98% nt identity) (P. brenneri type); and the sequence of one isolate of Pantoea spp. 4 showed the highest identity to that of L. adecarboxylata (95% nt identity).

Phylogeny of Pantoea isolates obtained in practical trials
A dendrogram was calculated using the partial rrs sequences of a length appropriate for the analysis (Fig. 2). A total of 29 rrs sequences were obtained in the present study. Based on the results of phylogenetic tree analyses on 29 strains, the Pantoea strains isolated in the present study were roughly divided into 2 groups. The first group mainly consisted of Pantoea species, including P. agglomerans, P. anthophila, P. brenneri, P. deleyi, P. vagans, and P. ananatis. The second  group consisted of P. stewartii, P. terrea, P. septica, P. punctata, and other Enterobacteriaceae (Fig. 2). Analyses using 16S rDNA enabled only the limited separation of strains within each Pantoea spp.
A total of 33 gyrB sequences were obtained. Based on the results of phylogenetic tree analyses for 33 strains, the Pantoea strains isolated in the present study were roughly divided into a number of groups (Fig. 3). In addition, P. agglomerans was further divided into three clades; clades 1 and 2 did not contain previously reported strains. The other clade formed a large group containing the P. agglomerans-type strain and already reported biocontrol strains (20). Similarly, P. vagans was also divided into two clades, in which clade 3 contained strain BD502, which was previously reported as a biocontrol strain (20), and clade 4 did not contain any previously reported strains. P. deleyi grouped with strain LMG24200, which was previously reported as an environmental strain (2). P. ananatis formed a large group in which all of the isolates were found to belong to Pantoea spp. 2. These results indicated that LOMAC medium is applicable to the isolation of Pantoea species exhibiting novel phylogenetic properties.

Discussion
We herein developed a new semi-selective agar medium and proposed a protocol for isolating Pantoea species. On LOMAC medium, Pantoea strains formed yellow colonies; however, some Gram-negative bacteria from environmental samples also formed yellow colonies. It is conceivable that under the same substrate availability conditions, strains other than Pantoea will form colonies of the same color. However, many Pantoea strains are known to produce a yellow pigment (5,8). When yellow colonies isolated on LOMAC medium were passaged on TSA medium, the majority of Pantoea strains in the present study produced a yellow pigment on TSA. This result indicates that the diagnostic accuracy of the procedure may be improved by eliminating bacteria that do not produce a yellow pigment on TSA. The efficient recovery of Pantoea strains on LOMAC medium suggests its applicability to investigations on the ecology of these species in the environment.
In practical trials, many Pantoea strains were isolated from plants. Pantoea is a plant-derived bacterium known to exist as an endophyte. As reported previously (21,22), the inside of plants is considered to be suitable for the survival of Pantoea species, such as P. vagans. Endophytic organisms, such as Pantoea, live inside plants without causing damage (17). In addition, endophytic Pantoea may promote plant growth by accelerating processes including nitrogen fixation, phosphate solubilization, siderophore secretion, and biocontrol (13,17). Different Pantoea species were detected in different parts of the same plant in the present study. For example, 6 strains of P. ananatis were isolated from the roots of crops (dicot) in trial 8, but were not detected in seeds. This result suggests that the role of parasitism differs among species and also that each Pantoea species may play a different role inside plants. Notably, P. deleyi was detected in trial 26 in the stem of a vegetable plant. Although limited information is currently available on P. deleyi, this species has been isolated from bacterial plaques and dead portions of eucalyptus (2).
However, there were no dead portions in the stems of the vegetables examined in this trial, and, therefore, the function of P. deleyi in these plants remains unclear. This species is predicted to be more closely related to P. vagans based on the results of gyrB phylogenetic tree analyses. Future studies will provide more information on this strain.
Pantoea strains were also isolated from 4 out of 24 spots of soil samples (16.7%) and from 15 out of 19 plant types (78.9%) with a large difference in detection rates. The population of Pantoea strains varies among crops, weeds, vegetables, fruits and soils in the environment in Japan.
Attempts to isolate Pantoea strains in 11 trials using samples from Nagano and 1 trial using samples from Hokkaido were unsuccessful because no colonies with a yellow color were observed on LOMAC medium. Among the non-yellow colonies, 7 were randomly selected and subjected to genetic analyses. As expected, gyrB sequencing revealed that these 7 colonies were not Pantoea species. In consideration of this result as well as recovery efficiency from soils, the protocol used in the present study is suitable for isolating Pantoea species.
In conclusion, we herein developed a new semi-selective medium known as LOMAC and established a protocol for isolating Pantoea species with high test efficiency. We detected Pantoea strains in samples of plants and soils from Japan using LOMAC even when Pantoea species were present at lower densities than non-target bacteria. Therefore, LOMAC medium enables the screening of Pantoea species from environmental sources and may be useful in future studies.