Genetic characteristics of Mycobacterium bovis BCG Thai Red Cross Society strain

BCG TRCS vaccine (lot No. FB03012), obtained from QSMI, was used for whole genomic sequencing. A working seed lot (BCG TRCS strain), which is used for production of BCG vaccine in Thailand, was characterized genetically.

Genomic DNA from BCG-TRCS vaccine (lot No. FB03012) was extracted using a genomic extraction kit (QIAamp DNA Mini Kit, Qiagen). Construction of the library and sequencing were performed by the Biochemistry Department, Medicine Faculty, Chulalongkorn University, Thailand. The library was prepared using a NEBNext Ultra DNA Library Prep Kit for Illumina. Whole genome sequencing and assembly were analyzed using a HiSeq 2000 Illumina system (New England Biolabs).

Genomic DNA from the working seed BCG (lot No. FB03012) was extracted using a genomic extraction kit (QIAamp DNA Mini Kit, Qiagen). PCR were performed with 2 primers RD16-F (5’-GGC TGG TGTT TCG TCA CTT C-3’), RD16-R (5’-ACA TTG GGA AAT CGC TGT TG-3’) [8]. PCR products were cloned using a pTG19-T vector (Vivantis, Malaysia) and then nucleotide sequences were determined by the 1^st Base DNA sequencing service (Seri Kembangan, Malaysia).

The complete genomic DNA of the BCG TRCS strain was submitted to GenBank and compared with the reference genomic DNA using the basic BLAST alignment tool (National Center for Biotechnology Information (NCBI) website at https://blast.ncbi.nlm.nih.gov/Blast.cgi#1051534617; U.S. National Institutes of Health). The genomic analysis tool was CLC Genomics Workbench 9 (CLC bio software, Qiagen). The complete genome of the BCG TRCS strain was compared to reference complete genome of M. bovis BCG Tokyo 172, GenBank: AP010918.

The working seed BCG TRCS strain was used as subject for conventional PCR amplification of the RD16 region with the primer set RD16-F and RD16- R. The PCR products showed 2 bands with different sizes (Figure 1).

Figure 1

Sequencing and alignment of the RD16 region in 2 different types of BCG were determined using CLC Genomics Workbench 9 (CLC bio software, Qiagen) and showed a 22 bp difference (Figure 2).

Figure 2

The results from sequencing and alignment of the RD16 region in 3 BCG strains are shown in Figure 3.

Figure 3

The genome of the BCG TRCS strain is 4,371,707 bp. The genomic sequences of the BCG TRCS strain and BCG Tokyo 172 strain were found to be homologous. The results from identification of 4,076 genes indicated 3,963 CDS; 3 genes for rRNA, 45 genes for tRNA, and 62 pseudogenes; genome sequence details were deposited in GenBank (accession no. CP014566). Comparison of whole genome sequence between the BCG TRCS strain (CP014566) and BCG Tokyo 172 (AP010918) by next-generation sequencing analysis found 23 different points in amino acids (Table 1).

Table 1

Different points of amino acids between the BCG TRCS strain (CP014566) and BCG Tokyo 172 (AP010918) found by next-generation sequencing.

The DNA alignment of the 16S rRNA gene showed several different points between BCG TRCS and BCG Tokyo 172 strains (Figure 4).

Figure 4

Single nucleotide polymorphisms uniquely represent the BCG Tokyo species and those of the BCG TRCS strain remain the same as for the Japanese species (Table 2).

Table 2

Comparison of single nucleotide polymorphisms between BCG TRCS strain (CP014566) and BCG Tokyo 172 (AP010918).

Figure 5 shows the results of ppsA gene alignment of BCG TRCS strain and BCG Tokyo type II (AB665170). The nucleotides at the position 275, 379, and 2415 are different.

Figure 5

Variable-number tandem repeat (VNTR) sequences are valuable markers for genotyping bacterial species. In the present study, 7 regions of difference (RDs) at the VNTR locus, which have shown polymorphism in the M. tuberculosis complex were used to compare the identities of the BCG species. The number from repeats from 7 RDs in the VNTR locus of the BCG TRCS strain was shown to be the same as that of BCG Tokyo 172 (Table 3).

Table 3

Seven regions of difference (RDs) in the variable-number tandem repeat (VNTR) locus of the BCG TRCS strain.

BCG vaccine was originally developed from one strain of M. bovis. Different methods of passage and storage have led to many substrains showing the phenotype. Such natural diversification, e.g. heterogeneity on duplication of the DU2 region in the BCG Danish substrains, commonly occurs in BCG substrains and can be increased by continuous passage without appropriate cloning because of natural mutations and selective pressure in culture media [11, 12]. BCG Tokyo 172 is one of 4 WHO vaccine reference strains [13, 14]. The Japanese BCG strain exhibits good protective efficacy with a low rate of unfavorable side effects [15]. Two types of variant strains, subpopulations I and II, are commonly found in the BCG Tokyo 172 vaccine preparation. They continue to coexist in subsequently produced lots with a decrease in the population of variant strain type II. BCG Tokyo 172 subpopulation types I and II are different, with a 22-bp deletion in the RD16 region. To our knowledge, there is no information on vaccine efficacy of the 2 different types [8].

The results of the RD16 region genotypes of the BCG TRCS strain from a working seed lot have shown a difference in the 2 PCR band product sizes, type I amplified 180 bp and type II 202 bp. PCR could identify type I with a 22-bp deletion and type II with the compete sequence in Figure 1 (lane 1). There were 22 bp differences between the 2 types (GAA GCT GAC CAG ACT GTT GCA C) as seen in Figure 2. These findings indicated that the BCG Tokyo 172- derived TRCS strain contained 2 types. PCR products of the RD16 BCG Tokyo type I subpopulation showed a characteristic 22-bp deletion, but type II was complete in this region. BCG Connaught and BCG Pasteur strain had a complete sequence in this region [8], while RD16 was missing in BCG Moreau [6, 8]. The TB H₃₇Rv PCR product in the RD16 region showed a band between BCG TRCS types I and II. By alignment of the RD16 region, we found that TB H₃₇Rv was similar to BCG Tokyo 172 type I and BCG TRCS type I. Therefore, TB H₃₇Rv also did not contain the 22 bp (Figure 3). A limitation of the present study is that we did not have BCG Tokyo strain, BCG Danish strain, or BCG Pasteur strains available to provide additional controls for PCR products.

Next-generation sequencing has indicated that the genome of BCG TRCS strain was 4,371,707 bp. DNA homology with BCG Tokyo 172 (AP010918) was 4,371,711 bp. Only 4 bp of DNA sequence and 23 4,371,711 bp. Only 4 bp of DNA sequence and 23 points of whole genome sequence were different between these BCG strains. VNTR sequences have emerged as valuable markers for the genotyping of several bacterial species, especially for genetically homogenous pathogens such as M. tuberculosis complexes [16]. Based upon the number of repeat polymorphisms within these tandemly arranged repetitive DNA sequences, many of these tandem loci display hypervariability, enabling their exploitation for strain typing in numerous bacterial species [17]. From Table 3, BCG TRCS strain matched BCG Tokyo 172 [5] showing 7 RDs retained their original sequence. PGL is antigenic and involved in host responses by acting as a cell wall component [8]. PGL can be found only in type I, but not in type II because of a ppsA gene mutation. BCG Tokyo type II (AB665170) had a single base insertion at nucleotide 379 (A base) and 2 single-base substitutions at nucleotides 275 (T→C) and 2415 (T→C). The length of ppsA gene CP014566 (BCG TRCS strain) was 5631 and AB665170 was 5632 (data not shown). The result in Figure 5 shows no mutation in the ppsA gene of BCG TRCS strain CP014566, indicating type I.

The difference between BCG TRCS strain (CP014566) and BCG Tokyo 172 (AP010918) might result from genetic change in PE-polymorphic GC- rich-repetitive sequence (PGRS) family protein, hypothetical protein, putative oxidoreductase, 16S rRNA, putative resuscitation promoting factor, putative ATP-dependent helicase, and sdhB (Table 1). The greatest genetic change position was for PE-PGRS family protein. PE-PGRS is a large family of typical proteins of pathogenic mycobacteria whose members are characterized by an N-terminal PE domain followed by a large Gly-Ala repeat-rich C-terminal domain [18]. The genes of the PE-PGRS family proteins are most often clustered in a region of the genome, often as overlapping genes. The proline- glutamic acid (PE) domain is responsible for the cellular localization of these proteins on bacterial cells [19]. The hypothetical proteins found 3 positions shown in Table 1 might be involved in virulence, detoxification, or adaptation; and proteins involved in intermediary metabolism and respiration [20]. The 16S rRNA gene is often used to study bacterial phylogeny and taxonomy. Table 1 shows there were the differences in positions 500–767 observed on the 16S rRNA DNA alignment between genes of the BCG Tokyo 172 (AP010918) and BCG TRCS strains (CP014566).

The result of 16S rRNA DNA alignment indicated some differences (Figure 4). The present study used next-generation sequencing, in which Q score >30 and genome coverage = 766×. This suggests the probability of an incorrect base is 1 in 1,000. The 16S rRNA region between BCG Tokyo 172 (AP010918) and BCG TRCS strain (CP014566) were apparently different. Similar occurrences were found in these genes, PE_PGRS16 (1093960–1095282), rpfE, PE_ PGRS43b, JTY_3396 and PE_PGRS57 (data not shown). We found that there are apparently many nucleotide differences between BCG-TRCS and BCG-Tokyo-172 strains in the 16S rRNA region. The degree of difference appears high given the 30-year history of the parent strain, and probable minimized passages of the seed lot. It is possible that these differences are the result of sequencing error. Sometimes new-generation sequencing can result in errors that depend on sample preparation or other mechanical factors. To be clear whether the result obtained is true or not, Sanger sequencing should be performed on the regions inconsistent with those of comparison. However, we were not able to perform Sanger sequencing or repeat the 16S rRNA DNA new-generation sequencing for want of more sample from this lot (No. FB03012). Use of more samples is warranted to confirm the sequencing.

The BCG Tokyo 172 preparation is mainly composed of the type I genotype, with known characteristics of this preparation, such as high viability and good heat stability [21]. The type I subpopulation shows a growth advantage over the type II subpopulation both on culture media and in mice. Although type II was still present in every preparation including the seed lot, the proportion of type II subpopulation decreases during passaging [22]. Type I produces PGL, but type II cannot because of the ppsA mutation. PGL deficient mutants show a phenotype with low virulence [23], and this might be a reason for type I always being the major component of the Tokyo 172 preparations [22]. The specimens from patients receiving BCG vaccination or BCG therapy for bladder cancer have only shown the genotype with the deletion in RD16 (type I) [24]. The present report does not explain why there are 2 variants of the BCG type in the vaccine. It is possible that some conditions of lot preparation, such as nutrient components and temperature in culture, might affect the proportions of certain subpopulations. Propagation of seed lots, subculturing to mass product, and freeze- drying techniques might generate mutations and give rise to the accumulation of genetic variants within the substrain  [11].

The BCG vaccine of BCG TRCS containing 2 types (I and II) from working seed lots found the proportion of type I to be more than type II in the PCR step. As the result of whole genome sequencing type I was found as the main population. Overall genetics of the BCG TRCS remain the same as BCG Tokyo 172 including the 7 RDs in ppsA, VNTR locus, and single nucleotide polymorphisms. Other genes appear to have been changed, and are apparently no longer the same as in BCG Tokyo 172, probably because of the storage environment, shelf life, preparation process, freeze drying, culture medium or natural mutation of the BCG Tokyo 172 species itself. Sequencing of the entire genome sequence of the BCG TRCS type II strain is planned for future work.