Molecular Taxonomic Evidence for Two Distinct Genotypes of Mycobacterium yongonense via Genome-Based Phylogenetic Analysis

Recently, we introduced a distinct Mycobacterium intracellulare INT-5 genotype, distantly related to other genotypes of M. intracellulare (INT-1 to -4). The aim of this study is to determine the exact taxonomic status of the M. intracellulare INT-5 genotype via genome-based phylogenetic analysis. To this end, genome sequences of the two INT-5 strains, MOTT-H4Y and MOTT-36Y were compared with M. intracellulare ATCC 13950T and Mycobacterium yongonense DSM 45126T. Our phylogenetic analysis based on complete genome sequences, multi-locus sequence typing (MLST) of 35 target genes, and single nucleotide polymorphism (SNP) analysis indicated that the two INT-5 strains were more closely related to M. yongonense DSM 45126T than the M. intracellulare strains. These results suggest their taxonomic transfer from M. intracellulare into M. yongonense. Finally, we selected 5 target genes (argH, dnaA, deaD, hsp65, and recF) and used SNPs for the identification of M. yongonese strains from other M. avium complex (MAC) strains. The application of the SNP analysis to 14 MAC clinical isolates enabled the selective identification of 4 M. yongonense clinical isolates from the other MACs. In conclusion, our genome-based phylogenetic analysis showed that the taxonomic status of two INT-5 strains, MOTT-H4Y and MOTT-36Y should be revised into M. yongonense. Our results also suggest that M. yongonense could be divided into 2 distinct genotypes (the Type I genotype with the M. parascrofulaceum rpoB gene and the Type II genotype with the M. intracellulare rpoB gene) depending on the presence of the lateral gene transfer of rpoB from M. parascrofulaceum.


Introduction
Members of the Mycobacterium avium complex (MAC) are the most important nontuberculous mycobacteria (NTM) in terms of clinical and epidemiological aspects [1]. Traditionally, MAC includes two species: M. avium and M. intracellulare [2][3][4]. In addition to these 2 species, recent advances in molecular taxonomy have fueled the identification of novel species within the MAC [5][6][7][8][9][10]. Our group introduced a novel species Mycobacterium yongonense, which was closely related to M. intracellulare, from a Korean patient with pulmonary symptoms [11]. Notably, M. yongonense possessed a distinct RNA polymerase gene (rpoB) sequence that was identical to M. parascrofulaceum, which is a distantly related scotochromogen, suggesting the acquisition of the rpoB gene via a potential lateral gene transfer (LGT) event [12,13]. Recently, M. yongonense strains causing pulmonary disease were also isolated from patients in Italy [14]. However, it should be noted that these strains harbored rpoB sequences that were almost identical to M. intracellulare and not M. parascrofulaceum, suggesting the possibility of the existence of another group of M. yongonense strains.
Our group reported that M. intracellulare-related strains from Korean patients showed genetic diversity. This diversity could be used to divide the strains into a total of five distinct groups (INT-1 to -5) via the molecular taxonomic approach using three independent chronometer molecules: hsp65, the internal transcribed spacer 1 (ITS-1) region and the 16S rRNA gene [15]. Of these genotypes, the INT-5 strains were distantly related to other genotypes of M. intracellulare (INT-1 to -4). We also introduced the complete genome sequences of two INT-5 strains, MOTT-H4Y and MOTT-36Y [16,17], showing that they were more closely related to the genome of M. yongonense DSM 45126 T than M. intracellulare ATCC 13950 T , despite they have rpoB sequences identical to M. intracellulare, but not M. parascrofulaceum. Furthermore, our recent study indicated that they harbored a novel insertion element (IS) sequence (ISMyo2) specific to M. yongonense [18]. Collectively, it suggests that MOTT-H4Y and MOTT-36Y might be variants of M. yongonense that were not subject to the rpoB LGT event from M. parascrofulaceum. Recently, it has been also reported that M. yongonense may be misidentified as one of the M. intracellulare strains [14,19]. Therefore, the establishment of consensus guidelines is needed for the exact species delineation between M. intracellulare and M. yongonense.
So, the aim of the current study is to determine the exact taxonomic status of the two INT-5 strains, MOTT-H4Y and MOTT-36Y with the M. intracellulare rpoB sequences but with genomic sequences closely related to M. yongonense via genome-based phylogenetic analysis.

Phylogenetic Analysis of Two INT-5 Strains, MOTT-H4Y and MOTT-36Y Based on the rpoB Gene Sequences and the Sequences of 35 Selected Target Genes
The taxonomic signature of M. yongonense was previously reported to be based on the rpoB gene sequence. The sequence of this gene is identical to the distantly related species M. parascrofulaceum, which enables the separation of the 2 closely related species M. intracellulare and M. yongonense [11,12]. Therefore, to obtain the exact taxonomic delineation of the two INT-5 strains we compared their taxonomic location by phylogenetic analysis based on the sequences of rpoB and 35 selected target genes.
The entire sequences of rpoB and the 35 selected genes were retrieved from the genome sequences of 6 mycobacterial strains   (Table 1). Detailed M. yongonense group-related SNP signatures are listed in Table 2.
In the case of rpoB gene, there was no M. yongonense group-related SNPs, however, rpoB gene of M. yongonense shared identical 151 SNPs with that of M. parascrofulaceum. A concatenated phylogenetic tree was constructed using the extracted SNP sequences. The tree showed that the two INT-5 strains were clustered into M. yongonense DSM 45126 T and separated from the other 3 M. intracellulare strains based on the phylogenetic analyses of the complete genome sequences and 35 concatenated gene sequences (Figs 1, 3B and 4A).

Application of the M. yongonense-Related SNP Analysis to MAC Clinical Isolates
To develop SNP analysis to enable the selective identification of M. yongonense strains from the MAC strains, five genes (argH, deaD, dnaA, hsp65 and recF) were selected that possessed a All the nucleotide positions were determined from Mycobacterium intracellulare ATCC 13950 T strain. a Bold characters represent M. yongonense group-related SNPs. doi:10.1371/journal.pone.0152703.t002

Discussion
In the present study, our phylogenetic analysis based on complete genome sequences, multilocus sequence typing (MLST) of 35 target genes, and single nucleotide polymorphism (SNP) analysis indicated that the two INT-5 strains, MOTT-H4Y and MOTT-36Y were more closely related to M. yongonense DSM 45126 T than the M. intracellulare strains. This finding suggests the presence of another distinct genotype in M. yongonense that may not have been subjected to the LGT event of rpoB from M. parascrofulaceum. Therefore, M. yongonense could be divided into 2 distinct genotypes: one with the M. parascrofulaceum rpoB gene and the other with the M. intracellulare rpoB gene, depending on the presence of the LGT event of rpoB from M. parascrofulaceum (Figs 1 and 3). Here, we proposed the former and the latter as the M. yongonense Type I and Type II genotypes, respectively.
To date, a total of 3 strains (M. yongonense DSM 45126 T , Asan 36912 and Asan 36527) belonging to the M. yongonense Type I genotype have been introduced via our 2 recent reports [11,12]. The Rhu strain used in this study was also identified as the Type I genotype by rpoB gene analysis (data not shown). In addition to MOTT-H4Y and MOTT-36Y, one additional strain (MOTT-68Y) used in this study was identified as the M. yongonense Type II genotype.
Although detailed taxonomic proof is needed, the M. yongonense strains recently isolated in Italy have the potential to be included in the M. yongonense Type II genotype.
LGT is the major mechanism by which bacteria can acquire genetic diversity, guaranteeing their survival under harsh environmental conditions [31,32]. However, it is generally accepted that mycobacteria are more resistant to LGT compared to other bacteria, possibly due to the unusually mycolic acid-rich cell wall structure and the relative scarcity of genetic elements such as plasmids and transposable elements [33][34][35]. Notably, because the M. yongonense strains were demonstrated to possess an rpoB gene that might have been laterally transferred from the distantly-related scotochromogenic species M. parascrofulaceum, these strains have gained increasing importance in the mycobacterial taxonomic fields. One of the noteworthy findings in this study is the identification of a novel genotype of M. yongonense without the rpoB gene from the LGT event in its genome. A genome comparison study between 3 mycobacterial groups [the M. yongonense Type I (subject to the LGT event) and Type II genotypes (without the LGT event) and M. parascrofulaceum (gene donor for LGT)] may provide novel insights into our understandings regarding mycobacterial LGT mechanisms.
In the present study, we developed an SNP analysis targeting 5 genes (argH, deaD, dnaA, hsp65 and recF) for the separation of M. yongonense from the closely related M. intracellulare strains. The concatenated 395-bp SNP-based phylogenetic analysis clearly separated 7 M. yongonense strains from 12 closely related M. intracellulare strains belonging to the INT-I and INT-2 genotypes, which were the first and the second most prevalent genotypes in Korean patients infected with M. intracellulare, respectively, with 83% bootstrap values (Fig 4A). This result suggests the feasibility of this assay for the selective identification of M. yongonense strains in clinical settings. Interestingly, this assay could not differentiate 4 Type I (DSM 45126 T , Asan36527, Asan 36912, and Rhu) and 3 Type II strains (MOTT-H4Y, MOTT-36Y and MOTT-68Y) (Fig 4B), suggesting the potential for gene exchanges by LGT events between the 2 genotypes. Notably, a total of 39 M. yongonense signature SNPs out of the 395 selected SNPs were found. These SNPs could be used for the development of M. yongonense-specific molecular diagnostic methods.
In conclusion, our genome-based phylogenetic analysis indicated that the taxonomic status of the two INT-5 strains, MOTT-H4Y and MOTT-36Y previously identified as M. intracellulare should be revised to M. yongonense. Taken together, M. yongonense could be divided into 2 distinct genotypes depending on the presence of the LGT event of rpoB from M. parascrofulaceum: the Type I genotype with the M. parascrofulaceum rpoB gene and the Type II genotype with the M. intracellulare rpoB gene. Additionally, we developed a novel SNP-based phylogenetic analysis to enable the taxonomic identification of M. yongonense clinical strains.
Supporting Information S1