Genomic expression catalogue of a global collection of BCG vaccine strains show evidence for highly diverged metabolic and cell-wall adaptations

Abstract

Although Bacillus Calmette-Guérin (BCG) vaccines against tuberculosis have been available for more than 90 years, their effectiveness has been hindered by variable protective efficacy and a lack of lasting memory responses. One factor contributing to this variability may be the diversity of the BCG strains that are used around the world, in part from genomic changes accumulated during vaccine production and their resulting differences in gene expression. We have compared the genomes and transcriptomes of a global collection of fourteen of the most widely used BCG strains at single base-pair resolution. We have also used quantitative proteomics to identify key differences in expression of proteins across five representative BCG strains of the four tandem duplication (DU) groups. We provide a comprehensive map of single nucleotide polymorphisms (SNPs), copy number variation and insertions and deletions (indels) across fourteen BCG strains. Genome-wide SNP characterization allowed the construction of a new and robust phylogenic genealogy of BCG strains. Transcriptional and proteomic profiling revealed a metabolic remodeling in BCG strains that may be reflected by altered immunogenicity and possibly vaccine efficacy. Together, these integrated-omic data represent the most comprehensive catalogue of genetic variation across a global collection of BCG strains.

Figure 1

(A) A circular representation of Bacillus Calmette-Guérin (BCG) vaccine sub-strains chromosomes. Distribution of single nucleotide polymorphisms (SNPs) in fourteen BCG strains compared to Mycobacterium bovis AF2122/97. Each coloured blob corresponds to the SNP density in a non-overlapping window of 10000 nucleotides, with a unique colour per sample allowing easy visual assessment of similarities between samples. The scale is shown in megabases in the circle. (B) Phylogenetic relationships among BCG strains. Maximum likelihood phylogeny tree based on 700 + variable common nucleotide positions across fourteen BCG genome sequences. The tree was rooted using M. bovis AF2122/97.

Figure 2. Revised genealogy of Bacillus Calmete-Guérin (BCG) vaccine strains.

The genealogy of BCG vaccine strains, displays the original virulent ancestor strain M. bovis and the subsequent series of genomic alteration including deletions of regions of difference (RD), and some strain-specific insertion (‘Ins’) and deletions (‘Δ’). Genes with both insertions and deletions is also shown (‘Indels’). Genetic markers identified in this work have been added to the scheme (squares bordered with a red solid line). Note that the RD nomenclature in previous work differs between references deletion of RD8 from BCG strains Connaught and Frappier corresponds to RD15 in reference. Variants shown in boxes are novel in this study. All variants shown were manually inspected by read mapping as well as by assembled genomes.

Figure 3. Transcriptional profile of a global collection of BCG strains: Gene expression heat maps of genes showing expression profiles involved in;

(A) Catabolism of fatty acids. (B) phenolpthiocerol dimycocerosate (PDIM). (C) mycolic acid synthesis. (D) Transcriptional regulatory. (E) Mammalian cell entry genes. (F) Immunogenic surface and secreted proteins. Genes were selected based on their annotation. The color scales represent log2-fold changes in gene expression (DESeq), using the M. bovis AF2122/97 strain as reference.

Figure 4. Pattern of increased and decreased expression in genes of BCG Japan and Danish compared to our own assembled genome of BCG Russia.

(A) Genes that are significantly down-regulated or up-regulated in BCG Japan compared to BCG Russia. Genes are categorized based on significant enrichment in biological processes of gene ontologies (p < 0.05). Left: total number of genes found to be down-regulated in each biological process category. Right: total number of genes found to be up-regulated in each biological process category. (B) Genes that are significantly down-regulated or up-regulated in BCG Danish compared to BCG Russia. Genes are categorized based on significant enrichment in biological processes of gene ontologies. Left: total number of genes found to be down-regulated in each biological process category. Right: total number of genes found to be up-regulated in each biological process category.

Figure 5. Heat map of differentially expressed proteins among five representative BCG strains compared to M. bovis.

Differential expression of proteins involved in biochemical pathways related to energy metabolism (A), or to biosynthesis pathways (B,C) are shown. The color scale indicates differential regulation as the Fold Change of the protein amount relative to the M. bovis 2122/97 reference strain. Up-regulation is indicated in red, down-regulation in blue. Star (*) indicates lack of correlation with RNA-seq data for that gene.