Detection of Rhodococcus fascians, the Causative Agent of Lily Fasciation in South Korea

Rhodococcus fascians is an important pathogen that infects various herbaceous perennials and reduces their economic value. In this study, we examined R. fascians isolates carrying a virulence gene from symptomatic lily plants grown in South Korea. Phylogenetic analysis using the nucleotide sequences of 16S rRNA, vicA, and fasD led to the classification of the isolates into four different strains of R. fascians. Inoculation of Nicotiana benthamiana with these isolates slowed root growth and resulted in symptoms of leafy gall. These findings elucidate the diversification of domestic pathogenic R. fascians and may lead to an accurate causal diagnosis to help reduce economic losses in the bulb market.


Introduction
The genus Rhodococcus contains Gram-positive aerobic bacterial species belonging to the phylum Actinobacteria. These bacteria can be isolated from water, soil, and marine habitats, including harsh ecological regions such as the Arctic, deserts, and heavily polluted areas [1,2]. Interest in Rhodococcus species has been increasing because of their biotechnological applications and potential in bioremediation [3,4]. Most Rhodococcus species grow epiphytically or endophytically in the roots or leaves of plants, are nonpathogenic, and are beneficial, aiding the host plant in environmental adaptation [5][6][7].
To date, only two members of the genus are known to be pathogenic: the plantpathogenic Rhodococcus fascians and livestock-pathogenic Rhodococcus equi [8][9][10]. Pathogenic R. fascians causes fasciation and leafy gall disease in a wide spectrum of plants, including 87 genera spanning 40 plant families [10][11][12]. The growth abnormalities and morphological alterations caused by this infection result in severe economic losses in ornamental horticulture industries in many countries. In addition to leafy gall, infected plants exhibit symptoms of witches' broom and other deformed growths [12], similar to those of plants with Agrobacterium infection or under abiotic stress [13,14].
Pathogenic R. fascians is often identified by its phenotypic characteristics in selection medium [15,16]; by molecular identification of virulence plasmids via PCR [17,18]; or by observation of disease symptoms in host plants, such as Nicotiana benthamiana [19]. Pathogenic R. fascians contains the linear plasmid pFiD188 with fas, att, and hyp virulence loci, and the cytokine production causes disease symptoms [20][21][22]. The only known virulence gene present on the chromosome of R. fascians is the vicA gene, which encodes malate synthase, an enzyme in the glyoxylate cycle [23]. However, R. fascians isolates without virulence genes are often cultured from symptomatic tissues [18,24].
In lily bulbs imported to South Korea, R. fascians has been detected as a result of strict domestic post-entry quarantine measures. A large proportion of these bulbs, the second most traded in South Korea, were thus destroyed, which became a social issue in South Korea [25]. Although not very large, the South Korean flower bulb market has been expanding steadily since 2016 and is regarded as a stable growing market among Dutch flower bulb exporters [26].
In this study, we aimed to identify the causative agent of fasciation in lily plants grown in Gangwon province, South Korea. The four identified R. fascians strains carried known virulence genes, such as fasD and vicA, and their infection caused disease symptoms in N. benthamiana. Overall, the accurate identification and characterization of R. fascians strains that induce lily fasciation will be helpful in the control and epidemiological investigation of this pathogen in South Korea.

Detection of R. fascians Virulence Gene in Symptomatic Lily Plants
To determine the causal agent of fasciation in lilies, eight hybrid symptomatic plants (YWS1 to YWS8), including Asian and Oriental lily varieties, were sampled from greenhouses and nurseries ( Figure 1, Supplementary Table S1). These plants exhibited fasciated flat stems, multiple-shoot proliferation, and stunted growth indicative of symptoms caused by R. fascians infection [12]. Figure 1B shows a plant with severe stunted growth and flat twisted stems. To verify infection by R. fascians, the total bacterial genomic DNA isolated from the plant samples was analyzed for the presence of the virulence gene fasD by PCR (Supplementary Table S2, Figure 1C). fasD is present in the fas operon in the pathogenic linear plasmid of R. fascians D188 [20] and encodes isopentenyl transferase, which is a major enzyme involved in cytokine synthesis that is widely used as a diagnostic marker for the pathogenicity of Rhodococcus species [17,18]. Sequencing of the amplified band revealed 100% identity with the fasD sequence of pathogenic R. fascians strain D188. These results suggest that the cause of the disease symptoms in these lilies was infection with pathogenic R. fascians.

Development of Genome-Based PCR Primers for Detection of Pathogenic R. fascians
Both plasmid-and chromosome-mediated virulence genes are involved in pathogenicity of R. fascians [19]. We detected the same fasD gene in all symptomatic samples and therefore tried to develop chromosome-based pathogenic markers for accurate screening of causative bacteria. We targeted vicA, the only virulence gene sent on the chromosome. However, there are reports that conventional vicA prim (VicA1497F and VicA1990R) [18] have problems in diagnosing phytopathogenic Rh coccus species [20]; therefore, we decided to develop new vicA-based diagnostic prim To this end, we obtained the full sequence of vicA from 50 Rhodococcus species inclu five pathogenic R. fascians strains from publicly available sources and identified vari regions that distinguish pathogenic species (Supplementary Figure S1), for which primers were designed (vicA44-F and vicA737R) (Supplementary Table S2). The de tion efficacy of these primers was tested via PCRs using genomic DNA of a subset o Rhodococcus species as templates (Figure 2A). DNA from purchased R. fascians (Lo Biochemica GmbH, Sauerlach, Germany) was used as a positive control. No ampli tion with these primers was observed for the DNA of the avirulent Rhodococcus spe

Development of Genome-Based PCR Primers for Detection of Pathogenic R. fascians
Both plasmid-and chromosome-mediated virulence genes are involved in the pathogenicity of R. fascians [19]. We detected the same fasD gene in all symptomatic lily samples and therefore tried to develop chromosome-based pathogenic markers for the accurate screening of causative bacteria. We targeted vicA, the only virulence gene present on the chromosome. However, there are reports that conventional vicA primers (VicA1497F and VicA1990R) [18] have problems in diagnosing phytopathogenic Rhodococcus species [20]; therefore, we decided to develop new vicA-based diagnostic primers. To this end, we obtained the full sequence of vicA from 50 Rhodococcus species including five pathogenic R. fascians strains from publicly available sources and identified variable regions that distinguish pathogenic species (Supplementary Figure S1), for which new primers were designed (vicA44-F and vicA737R) (Supplementary Table S2). The detection efficacy of these primers was tested via PCRs using genomic DNA of a subset of 10 Rhodococcus species as templates ( Figure 2A). DNA from purchased R. fascians (Loewe Biochemica GmbH, Sauerlach, Germany) was used as a positive control. No amplification with these primers was observed for the DNA of the avirulent Rhodococcus species, but amplification was detected for the positive control. Additionally, amplification was detected in symptomatic lily specimens but not in asymptomatic lilies ( Figure 2B). These results are consistent with fasD-based detection, demonstrating the reliability of the newly designed vicA-based primers for R. fascians detection. As more bacterial genomic sequences become available, additional regions may be revealed as suitable targets for the diagnosis of Rhodococcus species.
Pathogens 2021, 10, 241 4 of 11 but amplification was detected for the positive control. Additionally, amplification was detected in symptomatic lily specimens but not in asymptomatic lilies ( Figure 2B). These results are consistent with fasD-based detection, demonstrating the reliability of the newly designed vicA-based primers for R. fascians detection. As more bacterial genomic sequences become available, additional regions may be revealed as suitable targets for the diagnosis of Rhodococcus species.  Table S2). The experiments were repeated at least three times, and representative results are shown.  Table S2). The experiments were repeated at least three times, and representative results are shown.

Isolation of R. fascians from Symptomatic Lily Plants
To isolate the bacteria from plant tissues, serial dilutions of root extracts from symptomatic lily plants were plated on semi-selective medium for R. fascians. As R. fascians has an orange-colored pigment [12], we initially selected the colored bacterial colonies. Orange colonies were found in four out of eight plant samples with varied disease symptom severities. Ten orange colonies were picked from each plant sample and cultured, and PCR was used to determine the presence of vicA and fasD. For nucleotide sequence analysis of the fasD gene, a new primer was synthesized and used instead of the conventional fasD diagnostic primer (Supplementary Table S2). It was capable of amplifying a larger fragment and detecting both known fasD genes (Supplementary Figure S2). vicA-specific Pathogens 2021, 10, 241 5 of 11 amplification was detected in 19 of 40 samples, and fasD was also detected in 12 of these samples (Supplementary Figure S3). To investigate the genetic diversity of these 12 isolates and their relation to known pathogenic Rhodococcus species, vicA, fasD, and 16S rRNA genes were sequenced, and a phylogenetic analysis was performed (Supplementary Figure S4). BLAST search results showed that all isolates had the highest homology with pathogenic R. fascians strains in terms of 16s rRNA genes (≥ 99.4% identity) and fasD (100% identity) ( Table 1). In terms of vicA sequences, they showed 96.4-99.1% sequence similarity to D188. With nonpathogenic Rhodococcus KB6 as an outgroup [5], the phylogenetic analysis revealed that the 12 isolates were divided into four groups (Supplementary Figure S4). We also performed a phylogenetic analysis of the vicA sequences from strains representing each of these four groups and 50 known Rhodococcus species (Figure 3). All four strains belonged to the same lineage as the pathogenic Rhodococcus species, and they were closer to D188 with the linear virulence plasmid than to A25f carrying virulence genes on its chromosome [19,22,24]. Two strains, YWS3-1 and YWS8-2, belonged to the same branch as D188 (Figure 3). These results suggest that YWS isolates from symptomatic lilies are phytopathogenic R. fascians. Most plant pathogenic Rhodococcus have virulence plasmids; however, two exceptional strains, A21d2 and A25f, have been reported, which are presumed to have virulence genes in their chromosomes [24]. In particular, A21d2 lacks two-thirds of the fas operon, including fasD, which is replaced by a novel gene chimera in islands within the chromosome. Therefore, we investigated whether seven orange colonies that produced the pathogenic R. fascians-specific vicA amplicon, but not the fasD amplicon, are related to A21d2. A sequence comparison based on vicA showed that they all belong to the same clade as D188 (data not shown), suggesting that they lost their virulence plasmids during the screening process. ens 2021, 10, x FOR PEER REVIEW 6 of 13

Pathogenicity of YWS Isolates on Tobacco Plants
To verify the pathogenicity of the isolated strains using N. benthamiana [19], we selected isolates YWS1-1, YWS3-1, YWS4-1, and YWS8-2 as representative strains according to the vicA-based phylogenetic analysis (Figure 3). Infection of N. benthamiana seedlings with YWS3-1 and YWS8-2 significantly stunted root growth (p ≤ 0.05), whereas inoculations with isolates YWS1-1 and YWS4-1 did not alter growth compared with that of mock-treated control seedlings ( Figure 4A). Similarly, the meristems of tobacco plants inoculated with YWS3-1 and YWS8-2 exhibited the typical symptoms of leafy gall, including multiple proliferation of deformed leaves, foliar distortion, and generation of adventitious tissue ( Figure 4B). On the other hand, YWS1-1 and YWS4-1 induced less severe symptoms than the other two strains, and they delayed growth and weakened the dominance of the shoot apex compared with mock-inoculated control plants. These symptoms were more pronounced for YWS1-1 than for YWS4-1 ( Figure 4B). The bacteria were re-isolated from symptomatic N. benthamiana plant tissues, and vicA PCR analysis confirmed that they were the same bacteria as the inoculated bacteria (data not shown).  In conclusion, we demonstrated for the first time that the causative agent of the fasciated lily plant grown in Gangwon, South Korea, is phytopathogenic R. fascians. Publicly available whole genome sequence information of Rhodococcus species allowed us to develop diagnostic primers specific to pathogenic R. fascians by targeting the virulence gene vicA on the chromosome (Figure 2). In combination with the plasmid-based virulence marker fasD, we confirmed that the bacteria isolated from symptomatic lilies were pathogenic R. fascians. Furthermore, the isolated pathogenic R. fascians was classified into four strains based on the sequence diversity of the pathogenic vicA amplicon (Figure 3). Differences in the severity of disease symptoms caused by R. fascians D188 are related to the level of expression of the fas operon present in the virulent plasmid [20][21][22]. Although we have not yet observed fas transcripts expressed by the four isolates, the symptomatic difference in N. benthamiana may be due to differences in fas expression in these isolates. Another possibility is that the differences in symptom severity might be related to differences in the virulence genes present in their genomes. Indeed, nonpathogenic R. fascians can alter root structure, including the proliferation of root hairs and early lateral roots, without retarding root growth [20]. However, inoculations with YWS1-1 and YWS4-1 did not delay root growth, and we did not observe any deformation of roots ( Figure 4A). Therefore, we propose that the two isolates are D188 variants with the same virulence plasmid as D188, but with different pathogenicity due to differences in chromosomal virulence determinants rather than being nonpathogenic.
In conclusion, we demonstrated for the first time that the causative agent of the fasciated lily plant grown in Gangwon, South Korea, is phytopathogenic R. fascians. Publicly available whole genome sequence information of Rhodococcus species allowed us to develop diagnostic primers specific to pathogenic R. fascians by targeting the virulence gene vicA on the chromosome (Figure 2). In combination with the plasmid-based virulence marker fasD, we confirmed that the bacteria isolated from symptomatic lilies were pathogenic R. fascians. Furthermore, the isolated pathogenic R. fascians was classified into four strains based on the sequence diversity of the pathogenic vicA amplicon (Figure 3). This classification correlated with differences in pathogenicity of the isolates using N. benthamiana (Figure 4). These results will contribute to understanding the host range and pathogenic mechanisms of domestic R. fascians and will promote further efforts to control the pathogens that cause this disease.

Sample Collection and Bacterial Isolation
Eight symptomatic hybrid lilies (Lillium sp.) were collected from greenhouses or nurseries in Gangwon province, South Korea in May 2020 (Supplementary Table S1). The collected samples were treated with 1% sodium hypochlorite for 5 min and washed with sterile distilled water. Then, root samples were ground in half-strength phosphate-buffered saline containing 100 µg/mL cycloheximide, 40 µg/mL polymyxin B, and 0.4 µg/mL sodium azide. The extract was filtered through three layers of cheese cloth. The homogeneous extract was allowed to stand for 5 min. Supernatants were serially diluted and plated on modified D2 agar medium [15]. After incubation at 27 • C for 10 days, the orangecolored colonies were selected and restreaked for single-colony isolation. Bacterial isolates underwent a limited number of passages.

Molecular Identification of R. fascians
Orange-colored colonies were inoculated in nutrient broth (MBcell, Seoul, South Korea) and cultured at 27 • C for 2-3 days with shaking. Bacterial DNA was isolated using a Wizard genomic DNA purification kit (Promega Corp., Madison, WI) according to the manufacturer's instructions for Gram-positive bacteria. For bacterial identification, the 16S rRNA gene was amplified by PCR and analyzed after DNA sequencing. For further identification, the chromosomal virulence locus vicA (encoding malate synthase) and fasD (encoding isopentenyl transferase) were analyzed by PCR using R. fascians-specific primers. The conventional primer sets for fasD produced relatively short amplicons of ≈200 bp; therefore, a new primer set was designed using Primer Premier 6.00 to obtain a 573 bp PCR product (Supplementary Figure S2 and Table S2). To develop a new diagnostic vicA primer pair, the full sequence of vicA was obtained from 50 Rhodococcus species including five pathogenic R. fascians strains from publicly available sources, and multiple nucleotide sequence alignment was performed to identify regions that can be used to characterize pathogenic R. fascians strains (Supplementary Figure S1 and Table S2). The specificity of PCR primers was confirmed with total DNA isolated from 10 related Rhodococcus species (Figure 2). Primer sets for 16S rRNA, vicA, and fasD used in PCRs are described in Supplementary Table S2. Each 50 µL PCR mixture consisted of genomic DNA (50 ng), 2 µL forward primer (10 µM), and 2 µL reverse primer (10 µM) in 1× nTaq-tenuto PCR premix supplemented with 3'→5 proofreading activity (Enzynomics Co., Daejeon, South Korea). PCR was performed with the following cycling conditions: an initial denaturation at 95 • C for 3 min, 40 cycles of denaturation at 95 • C for 20 s, annealing at 55 • C for 30 s, and extension at 72 • C for 1 or 1.5 min. A final extension was performed at 72 • C for 10 min.

Phylogenetic Analysis
The PCR products of 16S rRNA, vicA, and fasD obtained from the bacterial isolates were fully sequenced by overlapping reads in forward and reverse directions using PCR primers and internal primers (785F, GGATTAGATACCCTGGTA and 800R, TACCAGGGTA-TCTAATCC for 16S rRNA). The resulting sequences were used as queries in BLAST searches (default search parameters) using the NCBI nr database. Based on a similarity of 16S rRNA (≥99%) with the reference sequences, the isolates were considered R. fascians (Table 1).
Phylogenetic trees were constructed with the sequences of 16S rRNA, vicA, and fasD using MEGA 7 software [27] and the neighbor-joining method [28]. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the maximum composite likelihood method. The sequences of reference R. fascians strains D188 and A25f were downloaded from the NCBI nr, nt, or wgs database (http://www.ncbi.nlm.nih.gov (accessed on 10 February 2021)) and used to construct the phylogenetic tree. The sequences from the Corynebacterium glutamicum strain AS019 (L27123) and nonpathogenic Rhodococcus species strain KB6 [5] were used as an outgroup for construction of the phylogenetic tree.

Plant Inoculation Assays
The inoculation of tobacco (Nicotiana benthamiana) seedlings was carried out as described previously [24]. Tobacco seeds were germinated on a half-strength MS agar medium containing 1% sucrose. After 3 days of incubation, the seedlings were inoculated with 3 µL R. fascians isolates (optical density at 600 nm (OD 600 ) of 0.5 in 10 mM MgCl 2 ) and vertically maintained at 26 • C under a 16 h light condition. Plant images were taken 7 days after inoculation. For the leafy gall assay of tobacco, apical meristems of 4-week-old plants were pinched using forceps and inoculated with 10 µL R. fascians suspension (OD 600 of 0.5 in 10 mM MgCl 2 ). Plant images were taken 3 weeks after inoculation. For a mock treatment, the plant was inoculated with 10 mM MgCl 2 . To isolate R. fascians from the infected plants, leafy galls or meristems were excised from plants and ground in halfstrength phosphate-buffered saline. Serial dilutions of extracts were plated on nutrient agar plates. After incubation for 3-5 days, orange-colored bacteria were analyzed by PCR using R. fascians-specific primers.

Statistical Analysis
Triplicates were performed with five specimens per experiment. Tukey's honestly significant difference test was used to determine significance using the Student's t-test (Microsoft Office Excel) and an ANOVA (SPSS v.18; IBM, Armonk, NY, USA). A p-value of <0.05 was considered statistically significant.
Supplementary Materials: The following are available online at https://www.mdpi.com/2076-081 7/10/2/241/s1, Figure S1: Multiple sequence alignment of the vicA genes of Rhodococcus species, Figure S2: Design of PCR primers for the fasD gene of R. fascians, Figure S3: Representative PCR-based detection of fasD, vicA, and 16S rRNA genes of bacterial isolates, Figure S4: Phylogenetic analysis of R. fascians YWS isolates from symptomatic lily plants, Table S1: Sampling place of lilies located in Gangwon province of South Korea, Table S2: List of oligonucleotide primers used in this study.