Two new clinical isolates from immigrants who were treated in Korea University Guro Hospital and Korea University Ansan Hospital were used for whole genome sequencing. The mycobacterial cultures were grown at 37℃ on Löwenstein-Jensen medium. Strains harvested at the early exponential phase were used for DNA extraction. All DNA extraction was performed as previously described⁹. At least one loop-full of cells (100 mg wet weight) were washed twice with TE (Tris-HCl pH 8.0, 10 mM; EDTA 1 mM) and an equal volume of saturated cesium chloride solution containing 1% Triton-X was added to it. Further purification was performed by the usual phenol-chloroform extraction and DNA was spooled out after adding 0.5 vol. of 5 M ammonium acetate and 0.75 vol. of isopropanol; it was washed with alcohol, partially dried and suspended in TE buffer. It was then treated with RNase A (50 µg/mL) for 30 minutes at 37℃, re-extracted with chloroformisoamyl alcohol, precipitated and finally re-suspended in TE buffer.

Whole-genome shotgun sequencing of the two strains was carried out using PacBio SMRT sequencing technology¹⁰ with ~150× depth. For assembly of the genomes, we applied the recently described hierarchical genome assembly process (HGAP) to the SMRT cells of sequencing data generated from an 8- to 10-kb insert library¹¹. Pre-assembly error correction was performed with HGAP of SMRT analysis version 2.3.0 (Pacific Biosciences, Menlo Park, CA, USA) using default parameters. Error-corrected reads were then assembled using Celera Assembler¹². To produce circular full-length sequence, each end section was resolved manually.

We compared the genome topology network (GTN) of each strain using a bootstrap topology tree and detected the locally collinear blocks of the conserved segments using a modified MUSCLE (multiple sequence alignment method with reduced time and space complexity) global alignment algorithm. We used the Mauve package for the identification and alignment of the conserved genomic DNA sequences of the nine M. tuberculosis strains⁸. Initially, local multiple alignments that had unique subsequences shared by two or more input genomes were selected, and ungapped extension was performed until the seed pattern no longer matched. Then, a progressive genome alignment according to a guide tree was built up after computing a pairwise genome content distance using the neighbor joining method and a pairwise breakpoint distance matrix. During progressive alignment, the breakpoint penalty according to the expected level of sequence divergence and the number of well-supported genomic rearrangements among the pair of input genomes were scaled. Anchor alignment using a global genome alignment algorithm was performed, and alignments that had unrelated sequences were ultimately rejected. The algorithm began with the initial set of matching regions (multiple maximal unique matches) represented as connected blocks. The matches were partitioned into a minimum set of collinear blocks. Each sequence of identically colored blocks represents a collinear set of matching regions. One connecting line is drawn per collinear block.

This study was approved by the Institutional Review Board (IRB) of Korea University Ansan Hospital (KUAS15157-001) and Korea University Guro Hospital (2017GR0301) with waivers of informed consents.

Table 1 presents the characteristics of two complete whole genome sequences from immigrants. The complete genome sequence of M. tuberculosis MTB_1 has been assigned the GenBank accession number CP020381.2. The M. tuberculosis MTB1 genome comprises one chromosome of 4,433,542 bp. In total, 4,306 genes comprising 4,255 coding sequences (CDS), three rRNA genes (5S, 16S, and 23S), 45 tRNAs, and three non-coding RNAs (ncRNAs) were annotated using NCBI Prokaryotic Genome Annotation Pipeline (PGAP; http://www.ncbi.nlm.nih.gov/genome/annotation_prok). The patient with the MTB_1 strain was a 39-year-old woman. She had recently emigrated from Shandong Province in China to South Korea to marry a Korean. Her chest computed tomogram showed a cavity in the left lung apex representing the reactivation of a previous TB infection.

The complete genome sequence of M. tuberculosis MTB_2 has been assigned the GenBank accession number CP022014.1. The M. tuberculosis MTB2 genome comprises one chromosome of 4,417,716 bp. In total, 4,290 genes comprising 4,239 CDS, 117 pseudo genes, three rRNA genes (5S, 16S, and 23S), 45 tRNAs, and three ncRNAs were annotated. The patient with MTB_2 strain was a 33-year-old man. He had moved to South Korea from the Yanbian Korean autonomous prefecture where most Korean-Chinese (Chosun race) reside, 2 years before the TB diagnosis was confirmed. His chest computed tomogram showed centrilobular nodules and branching opacities in both upper lung fields.

As shown in Table 2, the genomic sequences of seven strains (from Nos. 1 to 7) from native Chinese and South Korean individuals, which were obtained from a public website, were compared with MTB_1 and MTB_2 from the immigrants to South Korea. The Nos. 1 and 2 strains, which are known to be strains from native South Koreans, were isolated during an outbreak of pulmonary TB in high schools^13,14. The Nos. 3, 4, 5, and 6 strains were reported as strains from native China. The Nos. 3 and 4 strains were isolated from Chinese individuals living in Zunyi, which is located in southwest China¹⁵; the Nos. 5 and 6 strains were isolated from Chinese individuals living in Fujian Province, which is located in southeast China¹⁶; and the No. 7 strain was isolated from Chinese individuals living in Beijing, which is located in northwest China¹⁷. When the strains from the immigrants were analyzed, the MTB_2 strain was similar to the No. 1 and No. 2 strains isolated from native Koreans. On the other hand, the MTB_1 strain was similar to the strains isolated from the native Chinese individuals.

Figure 1 represents the Mauve alignment for nine M. tuberculosis strains, which shows collinear set of matched colored regions. And, in Figure 2, the phylogenetic trees exhibit similarities and differences in topology. The phylogenetic analysis of the GTN resulted in a rooted phylogenetic tree grouping the nine strains into two lineages: (1) strains from Chinese individuals and (2) strains from Korean individuals. However, even though the MTB_1 strain was grouped to a lineage from Chinese individuals (from the Nos. 3 to 7 strains), the MTB_1 strain was more similar to the Nos. 3 and 4 strains (Zunyi, China) than to the Nos. 5, 6, and 7 strains (Fujian Province and Beijing, China). On contrast, MTB_2 was grouped to a lineage from the Korean individuals (Nos. 1 and 2) that was different from the strains from the Chinese individuals, closer to the No. 2 strain than to the No. 1 strain.