Relaxed selection drives a noisy noncoding transcriptome in members of the Mycobacterium tuberculosis complex

Related species are often used to understand the molecular underpinning of virulence through examination of a shared set of biological features attributable to a core genome of orthologous genes. An important but insufficiently studied issue, however, is the extent to which the regulatory architectures are similarly conserved. A small number of studies have compared the primary transcriptomes of different bacterial species, but few have compared closely related species with clearly divergent evolutionary histories. We addressed the impact of differing modes of evolution within the genus Mycobacterium through comparison of the primary transcriptome of M. marinum with that of a closely related lineage, M. bovis. Both are thought to have evolved from an ancestral generalist species, with M. bovis and other members of the M. tuberculosis complex having subsequently undergone downsizing of their genomes during the transition to obligate pathogenicity. M. marinum, in contrast, has retained a large genome, appropriate for an environmental organism, and is a broad-host-range pathogen. We also examined changes over a shorter evolutionary time period through comparison of the primary transcriptome of M. bovis with that of another member of the M. tuberculosis complex (M. tuberculosis) which possesses an almost identical genome but maintains a distinct host preference. Importance: Our comparison of the transcriptional start site (TSS) maps of M. marinum and M. bovis uncovers a pillar of conserved promoters, noncoding RNA (NCRNA), and a genome-wide signal in the -35 promoter regions of both species. We identify evolutionarily conserved transcriptional attenuation and highlight its potential contribution to multidrug resistance mediated through the transcriptional regulator whiB7. We show that a species population history is reflected in its transcriptome and posit relaxed selection as the main driver of an abundance of canonical -10 promoter sites in M. bovis relative to M. marinum. It appears that transcriptome composition in mycobacteria is driven primarily by the availability of such sites and that their frequencies diverge significantly across the mycobacterial clade. Finally, through comparison of M. bovis and M. tuberculosis, we illustrate that single nucleotide polymorphism (SNP)-driven promoter differences likely underpin many of the transcriptional differences between M. tuberculosis complex lineages.