Comparative genomics of Mycobacterium africanum Lineage 5 and Lineage 6 from Ghana suggests distinct ecological niches

Abstract

Mycobacterium africanum (Maf) causes a substantial proportion of human tuberculosis in some countries of West Africa, but little is known on this pathogen. We compared the genomes of 253 Maf clinical isolates from Ghana, including N = 175 Lineage 5 (L5) and N = 78 Lineage 6 (L6). We found that the genomic diversity of L6 was higher than in L5 despite the smaller sample size. Regulatory proteins appeared to evolve neutrally in L5 but under purifying selection in L6. Even though over 90% of the human T cell epitopes were conserved in both lineages, L6 showed a higher ratio of non-synonymous to synonymous single nucleotide variation in these epitopes overall compared to L5. Of the 10% human T cell epitopes that were variable, most carried mutations that were lineage-specific. Our findings indicate that Maf L5 and L6 differ in some of their population genomic characteristics, possibly reflecting different selection pressures linked to distinct ecological niches.

Figure 1

Whole genome diversity of Maf Lineages (175 L5 and 78 L6 genomes). (a) Number of SNPs between Maf genomes and the hypothetical MTBC ancestor (the median fixed SNPs of L5 (934) is lower (W = 417, p-value < 2.2e-16) compared to L6 (1,039). (b) Pairwise SNPs between genomes within each lineage (the median of the pairwise SNPs is lower (W = 234, p-value < 2.2e-16) in L5 (212) compared to L6 (334). (c) Whole genome average nucleotide diversity (π) between L5 and L6 (the mean diversity of L5 (0.000076) is significantly (non-overlapping 95% confidence intervals) lower than L6 (0.000110). Error bars indicate 95% confidence intervals.

Figure 2

Phylogeny of Ghanaian Maf strains. The maximum likelihood phylogenetic tree of 253 Ghanaian Maf isolates is based on 11,027 variable positions. The tree was rooted on M. canettii and the confidence of nodes was assessed by bootstrapping 1000 pseudo replicates. Each lineage clade is colored according to the conventional MTBC lineage color codes.

Figure 3

Averaged nucleotide diversity (π) of Maf within genes of eight functional categories. epit – genes encoding human T cell epitopes, esmac – genes essential for growth in macrophages, intmedres – genes involved with intermediate metabolism and respiration, lipmet – genes involved with lipid metabolism, virdetad – genes involved with virulence, detoxification and adaptation, cwallproc - genes involed with cell wall and cell processes, regprot – genes encoding regulatory proteins and infopath – genes involved with information pathways. Error bars are indications of 95% confidence intervals.

Figure 4

Pairwise dN/dS of genes encoding human T cell epitopes and regulatory proteins in L5 and L6. Estimation of pairwise dN/dS of epitopes (a) and regulatory proteins (b) using the entire 147 L5 against the 67 L6 genomes. Estimation of pairwise dN/dS of epitopes (c) and regulatory proteins (d) using the mean dN/dS values of 10 random samples (size = 67, with replacement) of L5 against the 67 L6 genomes.

Figure 5

Number of human T cell epitopes with nonsynonymous SNPs (nsSNPs) stratified by Maf lineage. No significant difference (X-squared = 1.487, df = 1, p-value = 0.22) between the number of epitopes with nsSNPs among the 67 L6 genomes and L5 (mean values of 10 random samples of size = 67 with replacement).

Figure 6

Number of human T cell epitopes (a) and human T cell antigens (b) with amino acid substitutions stratified by Maf lineage. Green represents L6-specific mutant antigens or epitopes. Brown represents L5-specific mutant antigens or epitopes. Yellow represents antigens or epitopes mutated in both L5 and L6 but at different loci with different amino acid substitutions.

Figure 7

Pairwise dN/dS of sequences encoding human T cell epitopes (a) and genes encoding regulatory proteins (b) of L5 by patient ethnicity. L5 genomes from strains isolated from patients of the Ewe ethnicity (15 genomes) against, average values of 10 random samples of size 15 of L5 genomes of isolates from Non-Ewe patients.