Frequent transmission of the Mycobacterium tuberculosis Beijing lineage and positive selection for the EsxW Beijing variant in Vietnam

Abstract

To examine the transmission dynamics of Mycobacterium tuberculosis (Mtb) isolated from tuberculosis patients in Ho Chi Minh City, Vietnam, we sequenced the whole genomes of 1,635 isolates and compared these with 3,144 isolates from elsewhere. The data identify an underlying burden of disease caused by the endemic Mtb lineage 1 associated with the activation of long-term latent infection, and a threefold higher burden associated with the more recently introduced Beijing lineage and lineage 4 Mtb strains. We find that Beijing lineage Mtb is frequently transferred between Vietnam and other countries, and detect higher levels of transmission of Beijing lineage strains within this host population than the endemic lineage 1 Mtb. Screening for parallel evolution of Beijing lineage-associated SNPs in other Mtb lineages as a signal of positive selection, we identify an alteration in the ESX-5 type VII-secreted protein EsxW, which could potentially contribute to the enhanced transmission of Beijing lineage Mtb in Vietnamese and other host populations.

Figure 1. Circulating M. tuberculosis strains in HCMC are divided into multiple distinct lineages.

(a) Maximum-likelihood phylogeny of 1635 Mtb isolates collected from TB patients in HCMC, with backgrounds shaded by lineage. Exterior rings indicate presence of known antimicrobial resistance-associated mutations (coloured by drug, according to legend in top right). (b) Frequency distribution of lineages by month. (c) Frequency distribution of lineages by patient age group.

Figure 2. Properties of lineage subtrees for HCMC M. tuberculosis genomes.

(a) Distributions of terminal branch lengths for the 1635-strain phylogeny. (b) Mean subtree heights (y-axis; measured as mean node-to-tip distances for each subtree) vs subtree size (x-axis; number of descendant tips). Shaded region indicates standard error of the mean across subtrees of a given size; labels indicate lineage. (c) Stacked area plot showing number of clusters (y-axis) within each lineage (coloured as in panel a) identified using different maximum patristic distance thresholds to define clusters (x-axis).

Figure 3. Phylogenies of M. tuberculosis showing relationships between isolates from HCMC and other locations.

HCMC isolates are coloured grey, isolates from four other localised studies are coloured as in panel (d), other locations are shown in black. (a) Lineage 1 (n=675 genomes). (b) Lineage 2 (n=1871 genomes). (c) Lineage 4 (n=2066 genomes). (d) Number of transfers between Vietnam and other locations predicted by stochastic mapping of locations onto the Lineage 2 and 4 trees.

Figure 4. EsxW mutation at gene, mRNA, protein and heterodimer level.

(a) Variation in promoter region of esxW and other CFP10 paralogs extracted from H37Rv (Lineage 4). Sites are coloured by conservation, coordinates are relative to the start codon. (b) Variation in protein coding region. Tree shows maximum likelihood phylogeny inferred from amino acid sequences; box indicates QILSS proteins. Alignment of protein-coding DNA sequence is shown, coloured by conservation; arrow indicates site of homoplasic SNP in esxW, (A to G), resulting in Thr to Ala substitution at EsxW protein residue 2. (c-d) RNAseq results for esxW and QILSS/ESX-5 paralogs measured in 4 Mtb isolate pairs, each including a Lineage 1 or 4 EsxW-2Ala mutant and its genetically closest EsxW-2Thr relative, following 24h macrophage infection. mRNA levels were estimated from read counts uniquely mapping to the region -21 to +9 for each gene; normalized to total reads uniquely mapping to the locus encoding the ESX-5 machinery (and esxM) in each isolate. Boxes indicate interquartile range. (d) Difference between 2Ala mutant and 2Thr wildtype for each pair, relative to wildtype expression level. (e) Structural model of EsxW-EsxV heterodimer. EsxV is shown as a surface (grey) and EsxW as a ribbon (red) with key residues shown as labeled sticks. (f) Comparison of biophysical measurements of heterodimer binding affinity between wildtype and mutant EsxW. All binding curves were determined across four replicates by microscale thermophoresis (MST), and are represented as the mean ± standard deviation.